COMPOSITIONS AND METHODS OF USE THEREOF CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of priority to U.S. Provisional Application No. 63/411,672 filed September 30, 2022, which is hereby incorporated herein by reference in its entirety. STATEMENT OF GOVERNMENT SUPPORT This invention was made with government support under grant/contract number R01GM109988 awarded by the National Institutes of Health. The government has certain rights in the invention. BACKGROUND The cell surface receptors that bind to the extracellular matrix (ECM) have specificity for amino acid epitopes representative of the various extracellular matrix components. This presents a set of molecular markers that can differentiate subpopulations based on interactions with the extracellular matrix that can translate to measurable cell behaviors. Because the extracellular matrix receptors are on the surface of the cell, these receptors can be readily tagged with molecular probes for identification and classification. However, there remains a lack in technical approaches to isolate these subpopulations in a viable manner as a means to advance understanding of each subpopulation and hence propose and test treatment paradigms to inclusively target all subpopulations. The compositions and methods discussed herein address these and other needs. SUMMARY In accordance with the purposes of the disclosed compositions and methods as embodied and broadly described herein, the disclosed subject matter relates to compositions and methods of use thereof. For example, disclosed herein are compositions, devices, systems, and methods for detection of a target cell using a peptide-modified magnetic particle. For example, disclosed herein are assays comprising a peptide-modified magnetic particle comprising a particle having a plurality of peptides attached to a surface thereof. In some examples, each particle comprises a magnetic portion comprising a magnetic material. In some examples, each of the plurality of peptides comprises a capture portion configured to capture and bind with at least a first portion of a first target cell. In some examples, the capture portion is configured to mimic a first extracellular matrix component. In some examples, the particle has an average particle size of from 1 nanometer to 20 micrometers. In some examples, the particle has an average particle size of from 5 nanometers to 10 micrometers. In some examples, the particle has an average particle size of from 5 nanometers to 5 micrometers. In some examples, the particle has an average particle size of from 5 nanometers to 1 micrometer. In some examples, the particle is substantially spherical in shape. In some examples, the particle further comprises a plasmonic portion comprising a plasmonic material, such that the particle is a peptide-modified plasmonic magnetic particle. In some examples, the plasmonic portion of the particle is configured to enhance a Raman signal of at least a second portion of the first target cell bound to the capture portion of one or more of the plurality of peptides. In some examples, the particle is a core-shell particle, the magnetic portion comprising the core and the plasmonic portion comprising the shell, such that the plasmonic portion at least partially surrounds the magnetic portion. In some examples, the core has an average core size of from 1 nanometer to 20 micrometers. In some examples, the core has an average core size of from 1 nm to 10 micrometers. In some examples, the core has an average core size of from 1 nm to 5 micrometers. In some examples, the core has an average core size of from 1 nm to 1 micrometer. In some examples, the shell has an average thickness of from 1 nanometer to 5 micrometers. In some examples, the shell has an average thickness of from 1 nanometer to 1 micrometer. In some examples, the core is substantially spherical in shape. In some examples, the plasmonic material comprises a metal selected from the group consisting of Ru, Rh, Pd, Os, Ir, Pt, Au, Ag, Cu, Al, Mg, and combinations thereof. In some examples, the plasmonic material comprises a metal selected from the group consisting of Ru, Rh, Pd, Os, Ir, Pt, Au, Ag, Cu, Al, and combinations thereof. In some examples, the plasmonic material comprises a metal selected from the group consisting of Pt, Au, Ag, Cu, Al, and combination thereof. In some examples, the plasmonic material comprises a metal selected from the group consisting of Au, Ag, and combinations thereof. In some examples, the magnetic material comprises a metal selected from the group consisting of Fe, Mn, Ni, Gd, Cu, Co, V, Zn, and combinations thereof. In some examples, the magnetic material comprises a metal selected from the group consisting of Fe, Mn, Ni, Gd, and combination thereof. In some examples, the magnetic material comprises iron. In some examples, the magnetic material comprises an iron oxide. In some examples, the magnetic material comprises Fe ₃ O ₄ . In some examples, the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle is at least partially dispersed in a solvent. In some examples, the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle is disposed on a substrate. In some examples, the surface of the particle further comprises a plurality of ligands attached thereto. In some examples, the plurality of ligands comprise a second plurality of peptides. In some examples, the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle comprises a first population and a second population, the first population being modified with the first plurality of peptides and the second population being modified with the second plurality of peptides. In some examples, each of the second plurality of peptides has a binding portion configured to capture and bind to at least a first portion of a second target cell and/or at least a second portion of the first target cell. In some examples, each of the second plurality of peptides has a binding portion further configured to mimic a second extracellular matrix component, the second extracellular matrix component being different than the first extracellular matrix component. In some examples, the second target cell is different than the first target cell. In some examples, the binding portion is configured to selectively bind the first portion of the second target cell. In some examples, the binding portion is configured to selectively bind the second portion of the first target cell. In some examples, the capture portion of the first plurality of peptides is configured to selectively bind the first portion of the target cell. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently has an average length of from 3 to 60 amino acids. In some examples, the first extracellular matrix component and the second extracellular matrix component independently comprise a glycosaminoglycan or proteoglycan (e.g., a heparan sulfate, a chondroitin sulfate, a keratan sulfate, perlecan, aggrecan, betaglycan, agrin, neurocan, versican, brevican, decorin, biglycan, testican, bikunin, fibromodulin, lumican, or a combination thereof), a non-proteoglycan polysaccharide (e.g., a hyaluronic acid), a protein (e.g., a collagen, an elastin, a fibronectin, a vitronectin, a laminin, tenascin C, or a combination thereof), or a combination thereof. In some examples, the first extracellular matrix component and the second extracellular matrix component independently comprises a hyaluronic acid, a fibronectin, a laminin, a collagen, a glycosaminoglycan, tenascin C, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises integrin- binding mimetic peptide. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises a fibronectin mimetic peptide, a laminin mimetic peptide, a collagen mimetic peptide, a glycosaminoglycan mimetic peptide, a glycosaminoglycan mimetic polysaccharide, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises a fibronectin mimetic peptide, a laminin mimetic peptide, a collagen mimetic peptide, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides comprises at least 90% identity to RGD, IKVAV, YIGSR, YGYYGDALR, GFOGER, FYFDLR, xNYYSNS, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides comprises at least 90% identity to RGDS, RGDK, RGDF, IKVAV, YIGSR, YGYYGDALR, GROGER, FYFDLR, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides comprises at least 90% identity to cyclic RGDfC, CDPGYIGSR, or a combination thereof. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises a cell surface receptor. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises a surface protein, an integral protein, a transmembrane protein, or a combination thereof. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises an integrin, a cadherin, a tight junction protein, or a combination thereof. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises an integrin (e.g., one or more integrins). In some examples, the integrin comprises Į _V ȕ ₃ ^^Į _II ȕ ₃ ^^Į ₅ ȕ ₁ ^^Į ₃ ȕ ₁ ^^Į ₄ ȕ ₁ ^^Į ₆ ȕ ₁ ^^Į ₁ ȕ ₁ ^^Į ₂ ȕ ₁ , or a combination thereof. In some examples, the first target cell and/or the second target cell independently comprises a cancer cell. In some examples, the first target cell is a first subpopulation of cancer cells, and the second target cell is a second subpopulation of cancer cells, the first subpopulation and the second subpopulation being different. In some examples, the first target cell and/or the second target cell independently comprises a tumor cell. In some examples, the first target cell is a first subpopulation of tumor cells, and the second target cell is a second subpopulation of tumor cells, the first subpopulation and the second subpopulation being different. In some examples, the first target cell and/or the second target cell independently comprises a cell associated with brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. In some examples, the first target cell and/or the second target cell independently comprises a glioblastoma cell or a colorectal cancer cell. In some examples, the first target cell and/or the second target cell independently is a glioblastoma cell. In some examples, the first target cell is a first subpopulation of glioblastoma cells, and the second target cell is a second subpopulation of glioblastoma cells, the first subpopulation and the second subpopulation being different. In some examples, the first target cell and/or the second target cell independently is a colorectal cancer cell. In some examples, the first target cell is a first subpopulation of colorectal cancer cells, and the second target cell is a second subpopulation of colorectal cancer cells, the first subpopulation and the second subpopulation being different. Also disclosed herein are methods of making any of the assays disclosed herein. For example, the methods comprise making the peptide-modified magnetic particle or the peptide- modified plasmonic magnetic particle. In some examples, the method comprises contacting the particle with a plurality of peptides having a functional group configured to covalently or ionically bond to the particle. In some examples, the method further comprises making the particle. In some examples, the method further comprises making the first plurality of peptides having the functional group. Also disclosed herein are methods of use of any of the assays disclosed herein. For example, also disclosed herein are methods comprising contacting any of the assays disclosed herein with a liquid sample; subsequently collecting a surface enhanced Raman signal from the liquid sample and the assay; and processing the surface enhanced Raman signal to determine a property of the liquid sample. In some examples, the method further comprises applying a magnetic field to the liquid sample and the assay. Also disclosed herein are methods comprising contacting any of the assays disclosed herein with a liquid sample; and subsequently applying a magnetic field to the liquid sample and the assay. In some examples, the method further comprises collecting a surface enhanced Raman signal from the liquid sample and the assay; and processing the surface enhanced Raman signal to determine a property of the liquid sample. In some examples, applying a magnetic field comprises magnetic activated cellular sorting (MACS). In some examples, the method comprises repeatedly applying a magnetic field to the liquid sample and the assay. In some examples, the method comprises serial MACS, e.g. to separate different subpopulations of cells. In some examples, the property of the liquid sample comprises the presence of the target cell(s) in the liquid sample, the concentration of the target cell(s) in the liquid sample, the identity of the target cell(s), the subtype of the target cell(s), or a combination thereof. In some examples, the method further comprises diagnosing and/or monitoring a condition, a disease, or a disorder in a subject based on the property of the liquid sample. In some examples, the property of the liquid sample is indicative of the condition, the disease, or the disorder. In some examples, the disease comprises cancer. In some examples, the cancer comprises brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. In some examples, the cancer comprises glioblastoma or colorectal cancer. In some examples, the cancer comprises glioblastoma. In some examples, the method further comprises selecting a course of therapy for the subject based on the property of the liquid sample. In some examples, the method comprises cell subpopulation classification and/or isolation. Additional advantages of the disclosed compositions and methods will be set forth in part in the description which follows, and in part will be obvious from the description. The advantages of the disclosed compositions and methods will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosed compositions and methods, as claimed. The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims. BRIEF DESCRIPTION OF THE FIGURES The accompanying figures, which are incorporated in and constitute a part of this specification, illustrate several aspects of the disclosure, and together with the description, serve to explain the principles of the disclosure. Figure 1. Gene expression data demonstrates the 3D patient-derived tumor organoid (PTO) system better maintains integrin expression in Glioblastoma cells, while in 2D culture their expression decreases compared to patient-derived tumor organoid culture. Statistical significance: p<0.05. Figure 2A. Schematic illustration of functionalizing magnetic gold nanoparticles (NPs) with extracellular matrix peptide mimics (e.g., RGD, YIGSR, GFOGER, etc.) and imaging of peptide-nanoparticle binding to Glioblastoma derived cells. The peptide sequence can be altered to assay specificity. Figure 2B-Figure 2D. Multivariate analysis of SERS spectrum from particles bound to cells can be used to quantify the binding specificity based on the observed spectrum to the known (MCR-Ref) spectrum of the targeted receptor. The contribution of a specific receptor to the observed Raman signal is determined from multivariate analysis to determine probe specificity. Figure 2E-Figure 2F. Additional biophysical assays can be used to assess integrin binding, affinity, and selectivity, including Bradford assays (Figure 2E) and zeta-potential measurements (Figure 2F). Figure 3. Hyaluronic acid-based extracellular matrix hydrogels preserve genomic profiles of Glioblastoma biospecimen-derived cells. RNAseq hierarchical cluster analysis of originating tumor cells (Tu), organoids (Org), temozolomide-treated organoids (Org_T), and cells maintained in 2D tissue culture (Pla), from 6 human patients (A, C, E, F, G, and H). Tu, Org, and Org_T cluster together by patient, while Pla cultures cluster together, regardless of patient. Org/Org_T culture preserves the initial majority Glioblastoma subtype, while Pla promotes the mesenchymal subtype. Figure 4A. Bioprinting-based 3D invasion assay for increased throughput evaluation of Glioblastoma subpopulation invasiveness. a) Bioprinter printheads will be filled with a HA- based extracellular matrix hydrogel bio ink with fluorescently labeled Glioblastoma cell subpopulations or a fluorescently labeled version of the bioink. The bioprinter will create b) two adjacent regions with which invasion of labeled cells can be measured in relation to the interface of the fluorescent and non-fluorescent hydrogels. c) Demonstration of tumor cell invasion in this system. Figure 4B. Bioprinting-based 3D invasion assay for evaluation of GBM subpopulation invasiveness. a) Bioprinter printheads will be filled with a HA-based ECM hydrogel bio ink with GBM cell subpopulations in an ECM hydrogel precursor or a or the ECM hydrogel only. The bioprinter will create a 3D volume of ECM in each well, after which a second printhead will inject a dense cellularized tumor core within each ECM volume, and crosslink the 3D constructs. Also shown is an example of this invasion assay in practice using one highly invasive GBM subpopulation and one less invasive glioma subpopulation. Figure 4C. ATP quantification following temozolomide (TMZ) treatment of glioblastoma organoids generated from patient-derived glioblastoma cells sorted by fibronectin binding affinity using the fibronectin version of the extracellular matrix nanoparticles. There is a significant difference in viability (ATP activity is proportional to cell number) in organoids treated with 100 mM suggesting identification of a more resistant glioma subpopulation and a more responsive glioma subpopulation through sorting by fibronectin affinity. Statistical significance: * p < 0.05. Figure 4D. qRT-PCR analysis of gene expression shows dramatic differences between patient-derived glioblastoma cells sorted by fibronectin (FN) affinity suggesting extracellular matrix nanoparticles can identify true diverging subpopulations from a single glioma biospecimen. Figure 5. Schematic illustration of an example particle with a magnetic core, a gold shell, and a peptide cap. Figure 6. Schematic illustration of plasmonic effect. Figure 7. Schematic illustration of magnetic activated cellular sorting (MACS). Figure 8. SERS spectra taken of nonfunctionalized (bare) and fibronectin mimic- functionalized gold coated magnetic nanoparticles. The SERS results show a peak at 1001 cm ^-1 for the peptide mimic. Figure 9. Zeta potential measurements taken of nonfunctionalized (bare) and fibronectin mimic-functionalized gold coated magnetic nanoparticles at pH 4 and pH 7 which is above and below the isoelectric point of the peptide. Zeta Potential shows a positive increase as the peptide mimic is protonated. Figure 10. SERS spectra taken of fibronectin mimic-functionalized gold coated magnetic nanoparticles incubated with protein and the corresponding SERS peak assignments, illustrating fibronectin mimic protein binding. Figure 11. SERS spectra taken of nonfunctionalized (bare) and laminin mimic- functionalized gold coated magnetic nanoparticles. There was no discernable SERS signal for laminin peptide mimic. Figure 12. Zeta potential measurements taken of nonfunctionalized (bare) and laminin mimic-functionalized gold coated magnetic nanoparticles at pH 4 and pH 7, which is above and below the isoelectric point of the peptide. Zeta potential shows evidence of laminin peptide mimic functionalization. Figure 13. SERS spectra taken of laminin mimic-functionalized gold coated magnetic nanoparticles incubated with protein and the corresponding SERS peak assignments; a unique protein binding signal was observed after incubation of the with a target protein. Figure 14. Fibronectin Protein multivariate curve resolution (MCR) component generated from incubating protein with fibronectin mimic-functionalized magnetic nanoparticles in situ. Figure 15. Fibronectin negative selection scores plot. Figure 16. Fibronectin positive selection scores plot. Figure 17. SERS spectra taken of nonfunctionalized (bare) and Cyclic RGDfC functionalized gold coated magnetic nanoparticles. Appearance of 1001 cm ^-1 peak corresponds to phenylalanine. Figure 18. Zeta potential measurements taken of nonfunctionalized (bare) and Cyclic RGDfC functionalized gold coated magnetic nanoparticles at pH 4 and pH 7 which is above and below the isoelectric point of the peptide. Figure 19. Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles. Figure 20. Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles. A small plasmon shift (575 nm to 577 nm) is observed once peptide is bound to magnetic nanoparticles. Figure 21. SERS spectra taken of Cyclic RGDfC functionalized gold coated magnetic QDQRSDUWLFOHV^LQFXEDWHG^ZLWK^,QWHJULQ^Įvȕ3 and the corresponding SERS peak assignments. Figure 22^^%UDGIRUG^DVVD\^DEVRUEDQFH^VSHFWUD^RI^VWRFN^,QWHJULQ^Į _v ȕ ₃ (blue), Cyclic RGDfC Functionalized magnetic nanoparticles (red), Supernatant after magnetic separation ^\HOORZ^^^&\FOLF^5*'I&^ZLWK^,QWHJULQ^Įvȕ3 incubation remaining after magnetic separation. Figure 23^^,QWHJULQ^Į _v ȕ ₃ calibration curve using Bradford protein binding assay with corresponding LOD for the protein. Figure 24^^,QWHJULQ^Įvȕ3 calibration curve using Bradford protein binding assay with corresponding LOD for the protein. Figure 25. SERS spectra taken of nonfunctionalized (bare) and CDPGYIGSR functionalized gold coated magnetic nanoparticles. Figure 26. Zeta potential measurements taken of nonfunctionalized (bare) and CDPGYIGSR functionalized gold coated magnetic nanoparticles at pH 4 and pH 7 which is above and below the isoelectric point of the peptide. Figure 27. Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles. Figure 28. Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles. A small plasmon shift (575 nm to 577 nm) is observed once peptide is bound to magnetic nanoparticles. Figure 29. SERS spectra taken of Cyclic CDPGYIGSR functionalized gold coated magnetic nanopaUWLFOHV^LQFXEDWHG^ZLWK^,QWHJULQ^Į4ȕ1 and the corresponding SERS peak assignments. Figure 30. SERS spectra taken of CDPGYIGSR functionalized gold coated magnetic QDQRSDUWLFOHV^LQFXEDWHG^ZLWK^,QWHJULQ^Į4ȕ1 DQG^,QWHJULQ^Į6ȕ1. Figure 31. Microscope image of negatively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 32. Microscope image of negatively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 33. Microscope image of positively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 34. Microscope image of positively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 35. Microscope image of negatively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 36. Microscope image of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 37. Microscope image of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 38. Microscope image of negatively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 39. Microscope image of negatively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 40. Microscope image of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 41. Microscope image of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 42. Microscope image of negatively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 43. Microscope image of negatively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 44. Microscope image of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles. Figure 45. IntegriQ^Įvȕ3 Multivariate curve resolution component generated from LQFXEDWLQJ^,QWHJULQ^Įvȕ3 with Cyclic RGDfC functionalized magnetic nanoparticles in situ. Figure 46^^6FRUH^RI^,QWHJULQ^Į _v ȕ ₃ RQ^6(56^FHOO^PDS^FRUUHVSRQGLQJ^WKH^,QWHJULQ^Į _v ȕ ₃ multivariate curve resolution component for the protein on a positively sorted cell. The positive cell has the presence of higher scoring protein signal within the map (compare to Figure 47). Figure 47^^6FRUH^RI^,QWHJULQ^Į _v ȕ ₃ on SERS cell map corresponding WKH^,QWHJULQ^Į _v ȕ ₃ multivariate curve resolution component for the protein on both a negatively sorted cell. Figure 48. Absorbance measurements of bare and CDPGYIGSR functionalized magnetic nanoparticles. Figure 49. Zeta potential measurements taken of nonfunctionalized (bare) and CDPGYIGSR functionalized gold coated magnetic nanoparticles. Figure 50. SERS spectra taken of nonfunctionalized (bare) and CDPGYIGSR functionalized gold coated magnetic nanoparticles. Figure 51. Absorbance measurements of bare magnetic nanoparticles, CDPGYIGSR IXQFWLRQDOL]HG^PDJQHWLF^QDQRSDUWLFOHV^^^^^^^/^DQG^^^^^^L), and Cyclic RGDfC functionalized magnetic nanoparticles. Figure 52. Zeta potential measurements taken of nonfunctionalized (bare) magnetic nanoparticles, CDPGYIGSR fuQFWLRQDOL]HG^PDJQHWLF^QDQRSDUWLFOHV^^^^^^^/^DQG^^^^^^L), and Cyclic RGDfC functionalized magnetic nanoparticles. Figure 53. SERS spectra taken of nonfunctionalized (bare) magnetic nanoparticles, &'3*<,*65^IXQFWLRQDOL]HG^PDJQHWLF^QDQRSDUWLFOHV^^^^^^ ^/^DQG^^^^^^/^^^DQG^&\FOLF^5*'I&^ functionalized magnetic nanoparticles. Figure 54. CDPGYIGSR - Integrin A4B1 Bradford study. Figure 55. CDPGYIGSR - Integrin A4B1 separation study. Figure 56. Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles. Figure 57. SERS spectra taken of Cyclic RGDfC functionalized gold coated magnetic nanoparticles. Figure 58. SERS spectra taken of Cyclic RGDfC functionalized gold coated magnetic QDQRSDUWLFOHV^LQFXEDWHG^ZLWK^,QWHJULQ^Įvȕ3. Figure 59. RGDfC - Integrin Bradford assay absorbance. Figure 60. SERS spectra taken of Cyclic RGDfC functionalized particles incubated with ,QWHJULQ^Įvȕ3. Figure 61. SERS spectra of Cyclic RGDfC functionalized gold coated magnetic nanoparticles and citrate capped gold coated magnetic nanoparticles. Figure 62. Absorbance measurements of Cyclic RGDfC functionalized and citrate capped gold coated magnetic nanoparticles. Cyclic RGDfC Functionalization shown with absorption at the 260 nm region indicative of Phenylalanine as well as the 1003 cm ^-1 scattering peak. Figure 63. qRT-PCR analysis of several key genes important in Glioblastoma (CD44, EGFR, and Olig2) show dramatic differences between expression levels between cell subpopulations selected with FN peptide-functionalized beads and the remaining subpopulations. Figure 64. qRT-PCR analysis of several key genes important in Glioblastoma (CD44, EGFR, and Olig2) show dramatic differences between expression levels between cell subpopulations selected with FN peptide-functionalized beads and the remaining subpopulations. Figure 65. Analysis of additional genes by qRT-PCR show dramatic differences between expression levels between cell subpopulations selected with FN peptide-functionalized beads and the remaining subpopulations. Figure 66. Following FN-bead sorting, cells were used to create hyaluronic acid and gelatin-based 3D tumor constructs or “organoids” (thiolated hyaluronic and thiolated collagen, crosslinked with polyethylene glycol diacrylate in the presence of a photoinitiator under brief UV light pulses), which were subjected to drug screens with temozolomide (TMZ). “1 hr 1x NEG” indicates no bead was used, while “1 hr 1x POS” indicates the FN peptide-functionalize beads were used. While both organoid sets did ultimately respond to temozolomide at higher concentrations, at a moderate concentration (100 uM), the FN bead-selected cell subpopulation was significantly more resistant to the drug (statistical significance: * p < 0.5), suggesting true biological differences between these subpopulations. Figure 67. Schematic illustration of cancer cell targeting and SERS detection through cell surface receptors. Figure 68. Average SERS signals from baselined and normalized SERS spectra of Au nanoparticles and Cyclic RGDfC nanoparticles with shaded standard deviations. Figure 69. Average SERS signals from baselined and normalized SERS spectra of Au nanoparticles and CDPGYIGSR nanoparticles with shaded standard deviations. Figure 70. Zeta potential measurements of bare and cyclic RGDfC functionalized nanoparticles above and below the isoelectric point of the peptide. Figure 71. Zeta potential measurements of bare and CDPGYIGSR nanoparticles above and below the isoelectric point of the peptide. Figure 72. Normalized and average SERS spectra with standard deviations of Au nanoparticles functionalized with cyclo – RGDfC – ,QWHJULQ^Į _v ȕ ₃ , cyclo – RGDfC – Integrin Į ₅ ȕ ₁ , CDPGYIGSR – ,QWHJULQ^Į ₄ ȕ ₁ and CDPGYIGSR- ,QWHJULQ^Į ₆ ȕ ₁ and shaded standard deviations. Figure 73. Loadings for 4 component MCR model generated from peptide protein interaction on nanoparticles. Figure 74. Star coordinates plot of expected Integrin interactions with component 1 scores set at +X, component 2 scores set as +y, component 3 scores set as –x and component 4 scores set as –y for cyclic RGDfC. Figure 75. Star coordinates plot for CDPGYIGSR and expected binding integrins. Figure 76. Star Coordinates plot of expected Integrin interactions with component 1 scores set at +X, component 2 scores set as +y, component 3 scores set as –x and component 4 scores set as –y. Figure 77. Star coordinates plot for all off target interactions with peptides and integrins Figure 78. Percentage of cells sorted of SW480 (n=5) and SW 620 (n=6) colorectal cancer cells with standard deviation bars using Cyclic RGDfC functionalized nanoparticles. Figure 79. Percentage of SW 480 (n = 4) and SW 620 (n=7) colorectal cancer cells sorted with CDPGYIGSR functionalized nanoparticles. Figure 80. Dark field image of SW 480 Cyclic RGDfC Negative Sorted Cell . Figure 81. Dark field image of SW 480 CDPGYIGSR Negative Sorted Cell. Figure 82. Dark field image of SW 620 Cyclic RGDfC Negative Sorted Cell. Figure 83. Dark field image of SW 620 CDPGYIGSR Negative Sorted Cell. Figure 84. Dark field image of SW 480 Cyclic RGDfC Positive Sorted Cell. Figure 85. Dark field image of SW 480 CDPGYIGSR Positive Sorted Cell. Figure 86. Dark field image of SW 620 Cyclic RGDfC Positive Sorted Cell. Figure 87. Dark field image of SW 620 CDPGYIGSR Positive Sorted Cell. Figure 88. Extinction spectra normalized to extinction maximum of cyclic-RGDfC functionalized nanoparticles before and after functionalization. Extinction maximum is reported for each sample in the legend. In the extinction spectra, functionalization can further be confirmed by the extinction increase at the 200-210 nm wavelength region that can be attributed to absorbance of the peptide backbone. Figure 89. Extinction spectra normalized to extinction maximum of CDPGYIGSR functionalized nanoparticles before and after functionalization. Extinction maximum is reported for each sample in the legend. In the extinction spectra, functionalization can further be confirmed by the extinction increase at the 200-210 nm wavelength region that can be attributed to absorbance of the peptide backbone. Figure 90. Total Variance Explained for multiple MCR-ALS model systems and the corresponding fit percentage of each component. A 4 component model was determined to be optimal for modeling the 4 peptide, protein interactions. Figure 91. Stem Plot of integrated intensities for Cyclic RGDfC – Integrin Į ₄ ȕ ₁ interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. Figure 92. Stem Plot of integrated intensities for Cyclic RGDfC – Integrin Į ₅ ȕ ₁ interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal., Figure 93. Stem Plot of integrated intensities for Cyclic RGDfC – Integrin Į ₆ ȕ ₁ interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. Figure 94. Stem Plot of integrated intensities for Cyclic RGDfC – Integrin Įvȕ3 interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. Figure 95. Stem plots of integrated intensities for CDPGYIGSR – Integrin Į4ȕ1 interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. Figure 96. Stem plots of integrated intensities for CDPGYIGSR – Integrin Į5ȕ1 interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. Figure 97. Stem plots of integrated intensities for CDPGYIGSR – Integrin Į6ȕ1 interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. Figure 98. Stem plots of integrated intensities for CDPGYIGSR – Integrin Įvȕ3 interaction used in analysis of components. Red highlighted spectra were used for calculating the mean and standard and deviation for the spectra with no signal. DETAILED DESCRIPTION The compositions and methods described herein may be understood more readily by reference to the following detailed description of specific aspects of the disclosed subject matter and the Examples included therein. Before the present compositions and methods are disclosed and described, it is to be understood that the aspects described below are not limited to specific synthetic methods or specific reagents, as such may, of course, vary. It is also to be understood that the terminology used herein is for the purpose of describing particular aspects only and is not intended to be limiting. Also, throughout this specification, various publications are referenced. The disclosures of these publications in their entireties are hereby incorporated by reference into this application in order to more fully describe the state of the art to which the disclosed matter pertains. The references disclosed are also individually and specifically incorporated by reference herein for the material contained in them that is discussed in the sentence in which the reference is relied upon. In this specification and in the claims that follow, reference will be made to a number of terms, which shall be defined to have the following meanings. Throughout the description and claims of this specification the word “comprise” and other forms of the word, such as “comprising” and “comprises,” means including but not limited to, and is not intended to exclude, for example, other additives, components, integers, or steps. As used in the description and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a composition” includes mixtures of two or more such compositions, reference to “an agent” includes mixtures of two or more such agents, reference to “the component” includes mixtures of two or more such components, and the like. “Optional” or “optionally” means that the subsequently described event or circumstance can or cannot occur, and that the description includes instances where the event or circumstance occurs and instances where it does not. Ranges can be expressed herein as from “about” one particular value, and/or to “about” another particular value. By “about” is meant within 5% of the value, e.g., within 4, 3, 2, or 1% of the value. When such a range is expressed, another aspect includes from the one particular value and/or to the other particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” it will be understood that the particular value forms another aspect. It will be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint, and independently of the other endpoint. “Exemplary” means “an example of” and is not intended to convey an indication of a preferred or ideal embodiment. “Such as” is not used in a restrictive sense, but for explanatory purposes. Values can be expressed herein as an “average” value. “Average” generally refers to the statistical mean value. By “substantially” is meant within 5%, e.g., within 4%, 3%, 2%, or 1%. It is understood that throughout this specification the identifiers “first” and “second” are used solely to aid in distinguishing the various components and steps of the disclosed subject matter. The identifiers “first” and “second” are not intended to imply any particular order, amount, preference, or importance to the components or steps modified by these terms. The term “or combinations thereof” as used herein refers to all permutations and combinations of the listed items preceding the term. For example, “A, B, C, or combinations thereof” is intended to include at least one of: A, B, C, AB, AC, BC, or ABC, and if order is important in a particular context, also BA, CA, CB, CBA, BCA, ACB, BAC, or CAB. Continuing with this example, expressly included are combinations that contain repeats of one or more item or term, such as BB, AAA, AB, BBC, AAABCCCC, CBBAAA, CABABB, and so forth. The skilled artisan will understand that typically there is no limit on the number of items or terms in any combination, unless otherwise apparent from the context. As used herein, by a “subject” is meant an individual. Thus, the “subject” can include domesticated animals (e.g., cats, dogs, etc.), livestock (e.g., cattle, horses, pigs, sheep, goats, etc.), laboratory animals (e.g., mouse, rabbit, rat, guinea pig, etc.), and birds. “Subject” can also include a mammal, such as a primate or a human. Thus, the subject can be a human or veterinary patient. The term “patient” refers to a subject under the treatment of a clinician, e.g., physician. The term “inhibit” refers to a decrease in an activity, response, condition, disease, or other biological parameter. This can include but is not limited to the complete ablation of the activity, response, condition, or disease. This can also include, for example, a 10% reduction in the activity, response, condition, or disease as compared to the native or control level. Thus, the reduction can be a 10, 20, 30, 40, 50, 60, 70, 80, 90, 100%, or any amount of reduction in between as compared to native or control levels. By “reduce” or other forms of the word, such as “reducing” or “reduction,” is meant lowering of an event or characteristic (e.g., tumor growth). It is understood that this is typically in relation to some standard or expected value, in other words it is relative, but that it is not always necessary for the standard or relative value to be referred to. For example, “reduces tumor growth” means reducing the rate of growth of a tumor relative to a standard or a control. By “prevent” or other forms of the word, such as “preventing” or “prevention,” is meant to stop a particular event or characteristic, to stabilize or delay the development or progression of a particular event or characteristic, or to minimize the chances that a particular event or characteristic will occur. Prevent does not require comparison to a control as it is typically more absolute than, for example, reduce. As used herein, something could be reduced but not prevented, but something that is reduced could also be prevented. Likewise, something could be prevented but not reduced, but something that is prevented could also be reduced. It is understood that where reduce or prevent are used, unless specifically indicated otherwise, the use of the other word is also expressly disclosed. For example, the terms “prevent” or “suppress” can refer to a treatment that forestalls or slows the onset of a disease or condition or reduced the severity of the disease or condition. Thus, if a treatment can treat a disease in a subject having symptoms of the disease, it can also prevent or suppress that disease in a subject who has yet to suffer some or all of the symptoms. The term “treatment” refers to the medical management of a patient with the intent to cure, ameliorate, stabilize, or prevent a disease, pathological condition, or disorder. This term includes active treatment, that is, treatment directed specifically toward the improvement of a disease, pathological condition, or disorder, and also includes causal treatment, that is, treatment directed toward removal of the cause of the associated disease, pathological condition, or disorder. In addition, this term includes palliative treatment, that is, treatment designed for the relief of symptoms rather than the curing of the disease, pathological condition, or disorder; preventative treatment, that is, treatment directed to minimizing or partially or completely inhibiting the development of the associated disease, pathological condition, or disorder; and supportive treatment, that is, treatment employed to supplement another specific therapy directed toward the improvement of the associated disease, pathological condition, or disorder. Compositions and Assays Disclosed herein are compositions, devices, systems, and methods for detection of a target cell using a peptide-modified magnetic particle. For example, are assays comprising: a peptide-modified magnetic particle comprising a particle having a plurality of peptides attached to a surface thereof; wherein each particle comprises a magnetic portion comprising a magnetic material; wherein each of the plurality of peptides comprises a capture portion configured to capture and bind with at least a first portion of a first target cell; and wherein the capture portion is configured to mimic a first extracellular matrix component. As used herein, “a particle” and “the particle” are meant to include any number of particles. Thus, for example “a particle” includes one or more particles. In some embodiments, the particle can comprise a plurality of particles. The magnetic material can comprise any suitable material. In some examples, the magnetic material comprises a metal selected from the group consisting of Fe, Mn, Ni, Gd, Cu, Co, V, Zn, and combinations thereof. In some examples, the magnetic material comprises a metal selected from the group consisting of Fe, Mn, Ni, Gd, and combination thereof. In some examples, the magnetic material comprises iron. In some examples, the magnetic material comprises an iron oxide. In some examples, the magnetic material comprises Fe ₃ O ₄ . The particle can have an average particle size. “Average particle size” and “mean particle size” are used interchangeably herein, and generally refer to the statistical mean particle size of the particles in a population of particles. For example, the average particle size for a plurality of particles with a substantially spherical shape can comprise the average diameter of the plurality of particles. For a particle with a substantially spherical shape, the diameter of a particle can refer, for example, to the hydrodynamic diameter. As used herein, the hydrodynamic diameter of a particle can refer to the largest linear distance between two points on the surface of the particle. Mean particle size can be measured using methods known in the art, such as evaluation by scanning electron microscopy, transmission electron microscopy, and/or dynamic light scattering. The one or more particles can, for example, have an average particle size of 1 nanometer (nm) or more (e.g., 2 nm or more, 3 nm or more, 4 nm or more, 5 nm or more, 10 nm or more, 15 nm or more, 20 nm or more, 25 nm or more, 30 nm or more, 35 nm or more, 40 nm or more, 45 nm or more, 50 nm or more, 60 nm or more, 70 nm or more, 80 nm or more, 90 nm or more, 100 nm or more, 125 nm or more, 150 nm or more, 175 nm or more, 200 nm or more, 225 nm or more, 250 nm or more, 300 nm or more, 350 nm or more, 400 nm or more, 450 nm or more, 500 nm or more, 600 nm or more, 700 nm or more, 800 nm or more, 900 nm or more, 1 micrometer (micron) or more, 2 micrometers or more, 3 micrometers or more, 4 micrometers or more, 5 micrometers or more, 10 micrometers or more, or 15 micrometers or more). In some examples the one or more particles can have an average particle size of 20 micrometers or less (e.g., 15 micrometers or less, 10 micrometers or less, 5 micrometers or less, 4 micrometers or less, 3 micrometers or less, 2 micrometers or less, 1 micrometer or less, 900 nm or less, 800 nm or less, 700 nm or less, 600 nm or less, 500 nm or less, 450 nm or less, 400 nm or less, 350 nm or less, 300 nm or less, 250 nm or less, 225 nm or less, 200 nm or less, 175 nm or less, 150 nm or less, 125 nm or less, 100 nm or less, 90 nm or less, 80 nm or less, 70 nm or less, 60 nm or less, 50 nm or less, 45 nm or less, 40 nm or less, 35 nm or less, 30 nm or less, 25 nm or less, 20 nm or less, 15 nm or less, 10 nm or less, or 5 nm or less). The average particle size of the one or more particles can range from any of the minimum values described above to any of the maximum values described above. For example, the one or more particles can have an average particle size of from 1 nanometer to 20 micrometers (e.g., from 1 nm to 1 micrometer, from 1 micrometer to 20 micrometers, from 1 nm to 10 nm, from 10 nm to 100 nm, from 100 nm to 1 micrometer, from 1 micrometer to 10 micrometers, from 10 micrometers to 20 micrometers, from 1 nm to 10 micrometers, from 1 nm to 900 nm, from 1 nm to 750 nm, from 1 nm to 500 nm, from 1 nm to 250 nm, from 1 nm to 100 nm, from 5 nm to 20 micrometers, from 10 nm to 20 micrometers, from 100 nm to 20 micrometers, from 250 nm to 20 micrometers, from 500 nm to 20 micrometers, from 750 nm to 20 micrometers, from 5 nm to 15 micrometers, from 5 nm to 10 micrometers, from 10 nm to 10 micrometers, from 5 nm to 5 micrometers, or from 5 nm to 1 micrometer). In some examples, the one or more particles can have an average particle size of from 5 nm to 10 micrometers, from 5 nm to 5 micrometers, or from 5 nm to 1 micrometer. In some examples, the one or more particles can be substantially monodisperse. “Monodisperse” and “homogeneous size distribution,” as used herein, and generally describe a population of particles where all of the particles are the same or nearly the same size. As used herein, a monodisperse distribution refers to particle distributions in which 80% of the distribution (e.g., 85% of the distribution, 90% of the distribution, or 95% of the distribution) lies within 25% of the median particle size (e.g., within 20% of the median particle size, within 15% of the median particle size, within 10% of the median particle size, or within 5% of the median particle size). The one or more particles can comprise particles of any shape, such as a polyhedron (e.g., a platonic solid, a prism, a pyramid), a stellated polyhedron (e.g., a star), a cylinder, a hemicylinder, an elliptical cylinder, a hemi-elliptical cylinder, a sphere, a hemisphere, a cone, a semicone, etc. In some examples, the one or more particles can have a regular shape, an irregular shape, an isotropic shape, an anisotropic shape, or a combination thereof. In some examples, the one or more particles can have an isotropic shape or an anisotropic shape. In some examples, the one or more particles can have a shape that is substantially spherical, rod-like, pillar-like, or star- like. In some examples, the one or more particles are substantially spherical in shape. In some examples, the one or more peptide-modified magnetic particles can further comprise a plasmonic portion comprising a plasmonic material, such that the one or more particles are peptide-modified plasmonic magnetic particle(s). In some examples, the plasmonic portion of the particle is configured to enhance a Raman signal of at least a second portion of the first target cell bound to the capture portion of one or more of the plurality of peptides. The plasmonic material can comprise any suitable material. In some examples, the plasmonic material comprises a metal selected from the group consisting of Ru, Rh, Pd, Os, Ir, Pt, Au, Ag, Cu, Al, Mg, and combinations thereof. In some examples, the plasmonic material comprises a metal selected from the group consisting of Ru, Rh, Pd, Os, Ir, Pt, Au, Ag, Cu, Al, and combinations thereof. In some examples, the plasmonic material comprises a metal selected from the group consisting of Pt, Au, Ag, Cu, Al, and combination thereof. In some examples, the plasmonic material comprises a metal selected from the group consisting of Au, Ag, and combinations thereof. In some examples, each of the one or more particles is a core-shell particle. For example, wherein the magnetic portion comprises the core and the plasmonic portion comprises the shell, such that the plasmonic portion at least partially surrounds the magnetic portion of the particle. The core can have an average core size. “Average core size” and “mean core size” are used interchangeably herein, and generally refer to the statistical mean core size of the cores in a population of cores. For example, the average core size for a plurality of cores with a substantially spherical shape can comprise the average diameter of the plurality of cores. For a core with a substantially spherical shape, the diameter of a core can refer, for example, to the hydrodynamic diameter. As used herein, the hydrodynamic diameter of a core can refer to the largest linear distance between two points on the surface of the particle. Mean core size can be measured using methods known in the art, such as evaluation by scanning electron microscopy, transmission electron microscopy, and/or dynamic light scattering. The one or more cores can, for example, have an average core size of 1 nanometer (nm) or more (e.g., 2 nm or more, 3 nm or more, 4 nm or more, 5 nm or more, 10 nm or more, 15 nm or more, 20 nm or more, 25 nm or more, 30 nm or more, 35 nm or more, 40 nm or more, 45 nm or more, 50 nm or more, 60 nm or more, 70 nm or more, 80 nm or more, 90 nm or more, 100 nm or more, 125 nm or more, 150 nm or more, 175 nm or more, 200 nm or more, 225 nm or more, 250 nm or more, 300 nm or more, 350 nm or more, 400 nm or more, 450 nm or more, 500 nm or more, 600 nm or more, 700 nm or more, 800 nm or more, 900 nm or more, 1 micrometer (micron) or more, 2 micrometers or more, 3 micrometers or more, 4 micrometers or more, 5 micrometers or more, 10 micrometers or more, or 15 micrometers or more). In some examples the one or more cores can have an average core size of 20 micrometers or less (e.g., 15 micrometers or less, 10 micrometers or less, 5 micrometers or less, 4 micrometers or less, 3 micrometers or less, 2 micrometers or less, 1 micrometer or less, 900 nm or less, 800 nm or less, 700 nm or less, 600 nm or less, 500 nm or less, 450 nm or less, 400 nm or less, 350 nm or less, 300 nm or less, 250 nm or less, 225 nm or less, 200 nm or less, 175 nm or less, 150 nm or less, 125 nm or less, 100 nm or less, 90 nm or less, 80 nm or less, 70 nm or less, 60 nm or less, 50 nm or less, 45 nm or less, 40 nm or less, 35 nm or less, 30 nm or less, 25 nm or less, 20 nm or less, 15 nm or less, 10 nm or less, or 5 nm or less). The average core size of the one or more cores can range from any of the minimum values described above to any of the maximum values described above. For example, the one or more cores can have an average core size of from 1 nanometer to 20 micrometers (e.g., from 1 nm to 1 micrometer, from 1 micrometer to 20 micrometers, from 1 nm to 10 nm, from 10 nm to 100 nm, from 100 nm to 1 micrometer, from 1 micrometer to 10 micrometers, from 10 micrometers to 20 micrometers, from 1 nm to 10 micrometers, from 1 nm to 5 micrometers, from 1 nm to 900 nm, from 1 nm to 750 nm, from 1 nm to 500 nm, from 1 nm to 250 nm, from 1 nm to 100 nm, from 5 nm to 20 micrometers, from 10 nm to 20 micrometers, from 100 nm to 20 micrometers, from 250 nm to 20 micrometers, from 500 nm to 20 micrometers, from 750 nm to 20 micrometers, from 5 nm to 15 micrometers, or from 10 nm to 10 micrometers). In some examples, the one or more cores can have an average core size of from 1 nm to 10 micrometers, from 1 nm to 5 micrometers, or from 1 nm to 1 micrometer. In some examples, the one or more cores can be substantially monodisperse. “Monodisperse” and “homogeneous size distribution,” as used herein, and generally describe a population of cores where all of the cores are the same or nearly the same size. As used herein, a monodisperse distribution refers to core distributions in which 80% of the distribution (e.g., 85% of the distribution, 90% of the distribution, or 95% of the distribution) lies within 25% of the median core size (e.g., within 20% of the median core size, within 15% of the median core size, within 10% of the median core size, or within 5% of the core particle size). The one or more cores can comprise cores of any shape, such as a polyhedron (e.g., a platonic solid, a prism, a pyramid), a stellated polyhedron (e.g., a star), a cylinder, a hemicylinder, an elliptical cylinder, a hemi-elliptical cylinder, a sphere, a hemisphere, a cone, a semicone, etc. In some examples, the one or more cores can have a regular shape, an irregular shape, an isotropic shape, an anisotropic shape, or a combination thereof. In some examples, the one or more cores can have an isotropic shape or an anisotropic shape. In some examples, the one or more cores can have a shape that is substantially spherical, rod-like, pillar-like, or star- like. In some examples, the one or more cores are substantially spherical in shape. The shell can have an average thickness of 1 nanometer or more (e.g., 2 nm or more, 3 nm or more, 4 nm or more, 5 nm or more, 10 nm or more, 15 nm or more, 20 nm or more, 25 nm or more, 30 nm or more, 35 nm or more, 40 nm or more, 45 nm or more, 50 nm or more, 60 nm or more, 70 nm or more, 80 nm or more, 90 nm or more, 100 nm or more, 125 nm or more, 150 nm or more, 175 nm or more, 200 nm or more, 225 nm or more, 250 nm or more, 300 nm or more, 350 nm or more, 400 nm or more, 450 nm or more, 500 nm or more, 600 nm or more, 700 nm or more, 800 nm or more, 900 nm or more, 1 micrometer or more, 1.25 micrometers or more, 1.5 micrometers or more, 1.75 micrometers or more, 2 micrometers or more, 2.25 micrometers or more, 2.5 micrometers or more, 2.75 micrometers or more, 3 micrometers or more, 3.25 micrometers or more, 3.5 micrometers or more, 3.75 micrometers or more, 4 micrometers or more, 4.25 micrometers or more, 4.5 micrometers or more, or 4.75 micrometers or more). In some examples, the shell can have an average thickness of 5 micrometers (microns, μm) or less (e.g., 4.75 micrometers or less, 4.5 micrometers or less, 4.25 micrometers or less, 4 micrometers or less, 3.75 micrometers or less, 3.5 micrometers or less, 3.25 micrometers or less, 3 micrometers or less, 2.75 micrometers or less, 2.5 micrometers or less, 2.25 micrometers or less, 2 micrometers or less, 1.75 micrometers or less, 1.5 micrometers or less, 1.25 micrometers or less, 1 micrometer or less, 900 nm or less, 800 nm or less, 700 nm or less, 600 nm or less, 500 nm or less, 450 nm or less, 400 nm or less, 350 nm or less, 300 nm or less, 250 nm or less, 225 nm or less, 200 nm or less, 175 nm or less, 150 nm or less, 125 nm or less, 100 nm or less, 90 nm or less, 80 nm or less, 70 nm or less, 60 nm or less, 50 nm or less, 45 nm or less, 40 nm or less, 35 nm or less, 30 nm or less, 25 nm or less, 20 nm or less, 15 nm or less, 10 nm or less, or 5 nm or less). The average thickness of the shell can range from any of the minimum values described above to any of the maximum values described above. For example, the shell can have an average thickness of from 1 nanometer to 5 micrometers (e.g., from 1 nm to 2.5 μm, from 2.5 μm to 5 μm, from 1 nm to 1 μm, from 1 to 2 μm, from 2 to 3 μm, from 3 to 4 μm, from 4 to 5 μm, from 1 nm to 4 μm, from 1 nm to 3 μm, from 1 nm to 2 μm, from 1 nm to 1 μm, from 1 nm to 750 nm, from 1 nm to 500 nm, from 1 nm to 250 nm, from 1 nm to 100 nm, from 1 nm to 50 nm, from 1 nm to 25 nm, from 1 nm to 10 nm, from 5 nm to 5 μm, from 10 nm to 5 μm, from 25 nm to 5 μm, from 50 nm to 5 μm, from 75 nm to 5 μm, from 100 nm to 5 μm, from 250 nm to 5 μm, from 500 nm to 5 μm, from 1 to 5 μm, from 2 to 5 μm, from 3 to 5 μm, from 4 to 5 μm, from 1 nm to 4.5 μm, from 5 nm to 5 μm, from 5 nm to 4.5 μm, from 1 nm to 1 μm, from 1 nm to 500 nm, from 500 nm to 1 μm, from 1 nm to 200 nm, from 200 nm to 400 nm, from 400 nm to 600 nm, from 600 nm to 800 nm, from 800 nm to 1 μm, from 5 nm to 1 μm, from 10 nm to 1 μm, from 25 nm to 1 μm, from 50 nm to 1 μm, from 75 nm to 1 μm, from 100 nm to 1 μm, from 250 nm to 1 μm, from 1 nm to 900 nm, or from 5 nm to 900 nm). In some examples, the shell can have an average thickness of from 1 nanometer to 1 micrometer. In some examples, the peptide-modified magnetic particle and/or the peptide-modified plasmonic magnetic particle is at least partially dispersed in a solvent (e.g., a colloidal dispersion). The solvent can, for example, comprise water, ethylene glycol, polyethylene glycol, glycerol, alkane diol, ethanol, methanol, propanol, isopropanol, dimethyl sulfoxide (DMSO), acetonitrile, methylene chloride, or combinations thereof. In some examples, the solvent comprises water (e.g., an aqueous colloidal dispersion). In some examples, the peptide-modified magnetic particle and/or the peptide-modified plasmonic magnetic particle is disposed on a substrate. Examples of substrates include, but are not limited to, glass, quartz, silicon, silicon dioxide, nitrides (e.g., silicon nitride), polycarbonate, polydimethylsiloxane (PDMS), a cellulosic/lignin based substrate (e.g., wood, paper (such as chromatography paper), etc.), and combinations thereof. In some examples, the surface of the particle further comprises a plurality of ligands attached thereto. In some examples, the plurality of ligands comprise a second plurality of peptides. In some examples, the peptide-modified magnetic particle and/or the peptide-modified plasmonic magnetic particle comprises a first population and a second population, the first population being modified with the first plurality of peptides and the second population being modified with the second plurality of peptides (e.g., each population is modified with only the first peptide or the second peptide). In some examples, each of the second plurality of peptides has a binding portion configured to capture and bind to: at least a first portion of a second target cell and/or at least a second portion of the first target cell. The binding portion of the second plurality of peptides further being configured to mimic a second extracellular matrix component, the second extracellular matrix component being different than the first extracellular matrix component. In some examples, each of the second plurality of peptides has a binding portion configured to capture and bind to at least a first portion of a second target cell. The second target cell can be the same or different than the first target cell. In some examples, the second target cell is different than the first target cell. In some examples, each of the second plurality of peptides has a binding portion configured to capture and bind to at least a second portion of the first target cell. In some examples, the capture portion of each of the first plurality of peptides is configured to selectively bind the first portion of the target cell. The capture portion of each of the first plurality of peptides has a binding affinity for the first portion of the target cell. As used herein, “selectively bind” means that the binding affinity of the capture portion of each of the first plurality of peptides for the first portion of the target cell is greater than that of the binding affinity of the capture portion of each of the first plurality of peptides for another agent (e.g., another cell or cell portion) by a factor of 2 or more (e.g., 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, or 10 or more). The capture portion of each of the first plurality of peptides can, for example, have an average length of 3 amino acids or more (e.g., 4 amino acids or more, 5 amino acids or more, 6 amino acids or more, 7 amino acids or more, 8 amino acids or more, 9 amino acids or more, 10 amino acids or more, 15 amino acids or more, 20 amino acids or more, 25 amino acids or more, 30 amino acids or more, 35 amino acids or more, 40 amino acids or more, 45 amino acids or more, 50 amino acids or more, or 55 amino acids or more). In some examples, the capture portion of each of the first plurality of peptides can have an average length of 60 amino acids or less (e.g., 55 amino acids or less, 50 amino acids or less, 45 amino acids or less, 40 amino acids or less, 35 amino acids or less, 30 amino acids or less, 25 amino acids or less, 20 amino acids or less, 15 amino acids or less, 10 amino acids or less, 9 amino acids or less, 8 amino acids or less, 7 amino acids or less, 6 amino acids or less, or 5 amino acids or less). The average length of the capture portion of each of the first plurality of peptides can range from any of the minimum values described above to any of the maximum values described above. For example, the capture portion of each of the first plurality of peptides has an average length of from 3 to 60 amino acids (e.g., from 3 to 30 amino acids, from 30 to 60 amino acids, from 3 to 20 amino acids, from 20 to 40 amino acids, from 40 to 60 amino acids, from 3 to 10 amino acids, from 10 to 20 amino acids, from 20 to 30 amino acids, from 30 to 40 amino acids, from 40 to 50 amino acids, from 50 to 60 amino acids, from 5 to 60 amino acids, from 3 to 55 amino acids, from 5 to 55 amino acids, from 3 to 50 amino acids, from 3 to 40 amino acids, from 10 to 60 amino acids, or from 20 to 60 amino acids). In some examples, the binding portion of each of the second plurality of peptides is configured to selectively bind the first portion of the second target cell and/or to selectively bind the second portion of the first target cell. The binding portion of each of the second plurality of peptides has a binding affinity for the first portion of the second target cell or for the second portion of the first target cell. As used herein, “selectively bind” means that the binding affinity of the binding portion of each of the second plurality of peptides for the first portion of the second target cell or for the second portion of the first target cell is greater than that of the binding affinity of the binding portion of each of the second plurality of peptides for another agent (e.g., another cell or cell portion) by a factor of 2 or more (e.g., 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, or 10 or more). The binding portion of each of the second plurality of peptides can, for example, have an average length of 3 amino acids or more (e.g., 4 amino acids or more, 5 amino acids or more, 6 amino acids or more, 7 amino acids or more, 8 amino acids or more, 9 amino acids or more, 10 amino acids or more, 15 amino acids or more, 20 amino acids or more, 25 amino acids or more, 30 amino acids or more, 35 amino acids or more, 40 amino acids or more, 45 amino acids or more, 50 amino acids or more, or 55 amino acids or more). In some examples, the binding portion of each of the second plurality of peptides can have an average length of 60 amino acids or less (e.g., 55 amino acids or less, 50 amino acids or less, 45 amino acids or less, 40 amino acids or less, 35 amino acids or less, 30 amino acids or less, 25 amino acids or less, 20 amino acids or less, 15 amino acids or less, 10 amino acids or less, 9 amino acids or less, 8 amino acids or less, 7 amino acids or less, 6 amino acids or less, or 5 amino acids or less). The average length of the binding portion of each of the second plurality of peptides can range from any of the minimum values described above to any of the maximum values described above. For example, the binding portion of each of the second plurality of peptides has an average length of from 3 to 60 amino acids (e.g., from 3 to 30 amino acids, from 30 to 60 amino acids, from 3 to 20 amino acids, from 20 to 40 amino acids, from 40 to 60 amino acids, from 3 to 10 amino acids, from 10 to 20 amino acids, from 20 to 30 amino acids, from 30 to 40 amino acids, from 40 to 50 amino acids, from 50 to 60 amino acids, from 5 to 60 amino acids, from 3 to 55 amino acids, from 5 to 55 amino acids, from 3 to 50 amino acids, from 3 to 40 amino acids, from 10 to 60 amino acids, or from 20 to 60 amino acids). The first extracellular matrix component and the second extracellular matrix can comprise any suitable extracellular matrix component. In some examples, the first extracellular matrix component and the second extracellular matrix component independently comprise a glycosaminoglycan or proteoglycan (e.g., a heparan sulfate, a chondroitin sulfate, a keratan sulfate, perlecan, aggrecan, betaglycan, agrin, neurocan, versican, brevican, decorin, biglycan, testican, bikunin, fibromodulin, lumican, or a combination thereof), a non-proteoglycan polysaccharide (e.g., a hyaluronic acid), a protein (e.g., a collagen, an elastin, a fibronectin, a vitronectin, a laminin, tenascin C, or a combination thereof), or a combination thereof. In some examples, the first extracellular matrix component and the second extracellular matrix component independently comprises a hyaluronic acid, a fibronectin, a laminin, a collagen, a glycosaminoglycan, tenascin C, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides is independently a biomimetic peptide. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises an extracellular matrix component mimetic peptide. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises integrin- binding mimetic peptide. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises a fibronectin mimetic peptide, a laminin mimetic peptide, a collagen mimetic peptide, a glycosaminoglycan mimetic peptide, a glycosaminoglycan mimetic polysaccharide, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises a fibronectin mimetic peptide, a laminin mimetic peptide, a collagen mimetic peptide, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides comprises at least 90% (e.g., 90% or more, 91% or more, 92% or more, 93% or more, 94% or more, 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more) identity to RGD, IKVAV, YIGSR, YGYYGDALR, GFOGER, FYFDLR, xNYYSNS, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides comprises at least 90% (e.g., 90% or more, 91% or more, 92% or more, 93% or more, 94% or more, 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more) identity to RGDS, RGDK, RGDF, IKVAV, YIGSR, YGYYGDALR, GROGER, FYFDLR, or a combination thereof. In some examples, the capture portion of each of the first plurality of peptides comprises at least 90% (e.g., 90% or more, 91% or more, 92% or more, 93% or more, 94% or more, 95% or more, 96% or more, 97% or more, 98% or more, or 99% or more) identity to cyclic RGDfC, CDPGYIGSR, or a combination thereof. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises a cell surface receptor. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises a surface protein, an integral protein, a transmembrane protein, or a combination thereof. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises an integrin, a cadherin, a tight junction protein, or a combination thereof. In some examples, the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises an integrin (e.g., one or more integrins). In some examples, the integrin comprises ĮVȕ3^^ĮIIȕ3^^Į5ȕ1^^Į3ȕ1^^Į4ȕ1^^Į6ȕ1^^Į1ȕ1^^Į2 ȕ1, or a combination thereof. In some examples, the first plurality of peptides can comprise a peptide engineered to be specifically recognized by a protein of the target cell, such as a protein on the surface of the target cell. In some examples, the second plurality of peptides can comprise a peptide engineered to be specifically recognized by a protein of the second target cell or the first target cell, such as a protein on the surface of the second target cell or the first target cell. In some examples, the first target cell and/or the second target cell are independently indicative of a disease, a condition, and/or a disorder. In some examples, the disease comprises cancer. In some examples, the first target cell and/or the second target cell independently comprise a cancer cell. In some examples, the first target cell is a first subpopulation of cancer cells, and the second target cell is a second subpopulation of cancer cells, the first subpopulation and the second subpopulation being different. In some examples, the first target cell and/or the second target cell independently comprises a tumor cell. In some examples, the first target cell is a first subpopulation of tumor cells, and the second target cell is a second subpopulation of tumor cells, the first subpopulation and the second subpopulation being different. Examples of cancer types include bladder cancer, brain cancer, breast cancer, colorectal cancer, cervical cancer, gastrointestinal cancer, genitourinary cancer, head and neck cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, renal cancer, skin cancer, and testicular cancer. Further examples include cancer and/or tumors of the anus, bile duct, bone, bone marrow, bowel (including colon and rectum), eye, gall bladder, kidney, mouth, larynx, esophagus, stomach, testis, cervix, mesothelioma, neuroendocrine, penis, skin, spinal cord, thyroid, vagina, vulva, uterus, liver, muscle, blood cells (including lymphocytes and other immune system cells). Further examples of cancers treatable by the compounds and compositions described herein include carcinomas, Karposi’s sarcoma, melanoma, mesothelioma, soft tissue sarcoma, pancreatic cancer, lung cancer, leukemia (acute lymphoblastic, acute myeloid, chronic lymphocytic, chronic myeloid, and other), and lymphoma (Hodgkin’s and non-Hodgkin’s), and multiple myeloma. In some examples, the cancer and/or tumor comprises brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. In some examples, the cancer and/or tumor comprises glioblastoma or colorectal cancer. In some examples, the cancer and/or tumor comprises glioblastoma. In some examples, the first target cell and/or the second target cell independently comprises a cell associated with brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. In some examples, the first target cell and/or the second target cell independently comprises a glioblastoma cell or a colorectal cancer cell. In some examples, the first target cell and/or the second target cell is a glioblastoma cell. In some examples, the first target cell is a first subpopulation of glioblastoma cells, and the second target cell is a second subpopulation of glioblastoma cells, the first subpopulation and the second subpopulation being different. In some examples, the first target cell and/or the second target cell independently comprises a colorectal cancer cell. In some examples, the first target cell is a first subpopulation of colorectal cancer cells, and the second target cell is a second subpopulation of colorectal cancer cells, the first subpopulation and the second subpopulation being different. Methods of Making Also disclosed herein are methods of making any of the assays disclosed herein. For example, the methods can comprise making the peptide-modified magnetic particle and/or the peptide-modified plasmonic magnetic particle. In some examples, the methods comprise contacting the particle with a plurality of peptides having a functional group configured to covalently or ionically bond to the particle, such as thiol (-SH), carboxyl (-COOH), or amine (-NH) group. In some examples, the methods can further comprise making the particle. The particle can, for example, be made using standard techniques known in the art. In some examples, the methods can further comprise making the first plurality of peptides having the functional group, for example using standard techniques known in the art. Methods of Use Also disclosed herein are methods of use of any of the assays disclosed herein. For example, also disclosed herein are methods comprising contacting any of the assays disclosed herein with a liquid sample. The liquid sample can comprise any liquid sample of interest. By way of example the liquid sample can comprise a bodily fluid. "Bodily fluid", as used herein, refers to a fluid composition obtained from or located within a human or animal subject. Bodily fluids include, but are not limited to, urine, whole blood, blood plasma, serum, tears, semen, saliva, sputum, exhaled breath, nasal secretions, pharyngeal exudates, bronchoalveolar lavage, tracheal aspirations, interstitial fluid, lymph fluid, meningeal fluid, amniotic fluid, glandular fluid, feces, perspiration, mucous, vaginal or urethral secretion, cerebrospinal fluid, and transdermal exudate. Bodily fluid also includes experimentally separated fractions of all of the preceding solutions, as well as mixtures containing homogenized solid material, such as feces, tissues, and biopsy samples. In some examples, the methods can further comprise collecting the liquid sample. In some examples, the methods can further comprise purifying or treating the liquid sample before contacting the liquid sample with the assay. Purifying the liquid sample can, for example, comprise filtering, centrifuging, electrophoresis, extraction, or a combination thereof. Treating the liquid sample can, for example, comprise neutralization, buffer exchange, or a combination thereof. In some examples, the methods further comprise collecting a surface enhanced Raman signal from the liquid sample and the assay; and processing the surface enhanced Raman signal to determine a property of the liquid sample. In some examples, the methods further comprising applying a magnetic field to the liquid sample and the assay. Applying the magnetic field can, for example, comprise magnetic activated cellular sorting (MACS). In some examples, the method comprises repeatedly applying a magnetic field to the liquid sample and the assay. For example, the method can comprise serial MACS, e.g. to separate different subpopulations of cells. In some examples, the property of the liquid sample comprises the presence of the target cell(s) in the liquid sample, the concentration of the target cell(s) in the liquid sample, the identity of the target cell(s), the subtype of the target cell(s), or a combination thereof. In some examples, the methods further comprise diagnosing and/or monitoring a condition, a disease, or a disorder in a subject based on the property of the liquid sample. In some examples, the first target cell and/or the second target cell are independently indicative of a disease, a condition, and/or a disorder. In some examples, the property of the liquid sample is indicative of the condition, the disease, or the disorder. In some examples, the disease comprises cancer. In some examples, the first target cell and/or the second target cell independently comprise a cancer cell. In some examples, the first target cell and/or the second target cell independently comprises a tumor cell. Examples of cancer types include bladder cancer, brain cancer, breast cancer, colorectal cancer, cervical cancer, gastrointestinal cancer, genitourinary cancer, head and neck cancer, lung cancer, ovarian cancer, pancreatic cancer, prostate cancer, renal cancer, skin cancer, and testicular cancer. Further examples include cancer and/or tumors of the anus, bile duct, bone, bone marrow, bowel (including colon and rectum), eye, gall bladder, kidney, mouth, larynx, esophagus, stomach, testis, cervix, mesothelioma, neuroendocrine, penis, skin, spinal cord, thyroid, vagina, vulva, uterus, liver, muscle, blood cells (including lymphocytes and other immune system cells). Further examples of cancers treatable by the compounds and compositions described herein include carcinomas, Karposi’s sarcoma, melanoma, mesothelioma, soft tissue sarcoma, pancreatic cancer, lung cancer, leukemia (acute lymphoblastic, acute myeloid, chronic lymphocytic, chronic myeloid, and other), and lymphoma (Hodgkin’s and non-Hodgkin’s), and multiple myeloma. In some examples, the cancer and/or tumor comprises brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. In some examples, the cancer and/or tumor comprises glioblastoma or colorectal cancer. In some examples, the cancer and/or tumor comprises glioblastoma. In some examples, the first target cell and/or the second target cell independently comprises a cell associated with brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. In some examples, the first target cell and/or the second target cell independently comprises a glioblastoma cell or a colorectal cancer cell. In some examples, the first target cell and/or the second target cell is a glioblastoma cell. In some examples, the methods further comprise selecting a course of therapy for the subject based on the property of the liquid sample. In some examples, the methods comprise cell subpopulation classification and/or isolation. A number of embodiments of the invention have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the invention. Accordingly, other embodiments are within the scope of the following claims. The examples below are intended to further illustrate certain aspects of the systems and methods described herein, and are not intended to limit the scope of the claims. EXAMPLES The following examples are set forth below to illustrate the methods and results according to the disclosed subject matter. These examples are not intended to be inclusive of all aspects of the subject matter disclosed herein, but rather to illustrate representative methods and results. These examples are not intended to exclude equivalents and variations of the present invention which are apparent to one skilled in the art. Efforts have been made to ensure accuracy with respect to numbers (e.g., amounts, temperature, etc.) but some errors and deviations should be accounted for. Unless indicated otherwise, parts are parts by weight, temperature is in °C or is at ambient temperature, and pressure is at or near atmospheric. There are numerous variations and combinations of measurement conditions, e.g., component concentrations, temperatures, pressures and other measurement ranges and conditions that can be used to optimize the described process. Example 1 – Extracellular Matrix Nanoparticle Probes for Improved Tumor Subpopulation Classification and Isolation Described herein is the development of nanotechnology to improve classification and treatment of solid tumors (e.g., glioblastoma) by characterizing intratumoral heterogeneity based on extracellular matrix binding interactions. The nanoparticle probe technology, which can enable spectroscopic imaging and cell sorting, can improve classification and treatment of cancer, such as glioblastoma. The heterogenous nature of many tumors, especially malignant tumors, has long been appreciated histologically and more recent transcriptomic studies have greatly extended this view. This heterogeneity is representative of different cellular subpopulations, each with different biological behaviors and differential responses to therapeutic intervention. Hence, any given subpopulation may not effectively respond to a given treatment and therefore drive recurrence and poor clinical outcomes. Additionally, a different subpopulation in that same tumor may harbor a more invasive phenotype and evade surgical resection and/or contribute to metastasis. Central to dictating cellular phenotypes are the interactions of tumor cells with signals within the tumor microenvironment (TME), such as interactions with the extracellular matrix (ECM). The cell surface receptors on tumor cells that bind to the extracellular matrix have specificity for amino acid epitopes representative of the various extracellular matrix components. This presents a set of molecular markers that can differentiate subpopulations based on interactions with the extracellular matrix that translate directly to measurable cell behaviors associated with malignancy. Because the extracellular matrix receptors are on the surface of the cell, these receptors can be readily tagged with molecular probes for identification and classification. However, there remains a major lack in technical approaches to isolate these subpopulations in a viable manner as a means to advance the understanding of each subpopulation and hence propose and test treatment paradigms to inclusively target all subpopulations. Described herein is the development of a probe set, based on integrin-binding extracellular matrix peptide sequences covalently bound to magnetic nanoparticles that are amenable to surface enhanced Raman spectroscopy (SERS) and magnetic cellular sorting, thus leveraging the interactions between tumor cells and their physical tumor microenvironment, for the characterization and sorting of tumor cell subpopulations. A focus is on glioblastoma (GBM), an aggressive, terminal brain tumor with known extensive tumor cell and extracellular matrix heterogeneity. Specific extracellular matrix binding site-Glioblastoma cell interactions can serve as biomarkers that can be leveraged to categorize Glioblastoma subpopulations. A first set of experiments focused on developing and expanding the probe portfolio to account for three extracellular matrix adhesion protein peptide sequences (collagen Type I, fibronectin, and laminin), and validation of quantification via protein affinity assays, Raman imaging, and particle tracking experiments. A second set of experiments can validate this probe platform further, using a biobank of Glioblastoma patient-specific organoids in i) a 3D invasion assay, ii) chemotherapy response assays iii) a stemness assay, and iv) RNA sequencing to evaluate extracellular matrix probe-based analysis and sorting in terms of generation of statistically distinct subpopulations. Glioblastoma as a model can validate an initial panel of peptide- nanoparticle probes and establish a workflow that can be used to rapidly identify and sort cells from preclinical and clinical samples in multiple tumor types based on functional extracellular matrix affinities and lay the foundation for expansion of the probe set for additional components of the extracellular matrix. The tumor microenvironment (TME) directly influences malignant tumor cellular phenotypes such as invasiveness, treatment resistance, and proliferative state, but technical methodologies to isolate these subpopulations specifically based on tumor cell-tumor microenvironment interactions are limited. Invasive and aggressive tumors interact with the extracellular matrix (ECM) in the tumor microenvironment via cell surface receptors, and these interactions and extracellular matrix compositions vary locally. Here, glioblastoma, the most malignant and heterogeneous brain tumor, is utilized to isolate live cell subpopulations via their extracellular matrix interactions using a set of nanoparticle-peptide probes representative of specific extracellular matrix adhesion proteins that are amenable to quantitative analysis by Raman imaging and magnetic bead sorting for downstream applications. The advent of single cell sequencing has brought further resolution of the extensive intratumoral cellular heterogeneity of cancer at the transcriptomic level. However, there remains a major lack in technical approaches to isolate these subpopulations in a viable manner. Such methodology is essential for analyzing functional cell phenotypes correlated to malignancy and how treatment may augment or reduce such phenotypes. A probe set can be developed to leverage the interactions between tumor cells and their physical tumor microenvironment, for the characterization of tumor cell populations. Invasive and aggressive tumors – those with low prognoses – interact with the extracellular matrix (ECM) and these interactions instruct malignant cell phenotypes. The cell surface receptors that bind to the extracellular matrix have specificity for amino acid epitopes representative of the various extracellular matrix components: RGD for fibronectin, IKVAV and YIGSR for laminin, others (collagens, etc.). This presents a set of molecular markers that can differentiate subpopulations based on interactions with the extracellular matrix that translate directly to measurable cell behaviors associated with malignancy. Because the extracellular matrix receptors are on the surface of the cell, these receptors can be readily tagged with molecular probes for identification and classification. Glioblastoma (GBM; isocitrate dehydrogenase [IDH]-wild-type) is an excellent tumor model for the exploration of this approach. Glioblastoma is the most common primary malignant brain tumor with dismally low long-term survival and nearly universal recurrence within a year. Curative treatments remain elusive in major part due to the extensive intratumoral heterogeneity and aggressive and invasive nature of these tumors. Glioblastoma is a prime example whereby both bulk and single cell transcriptomic analyses have highlighted expression-based subtypes. However, molecular subtyping has largely failed to yield any transformative improvement to clinical outcomes. Interrogation of varying cellular phenotypes in Glioblastoma has so far focused on isolating subpopulations using antibody-based sorting of cell surface markers. However, a limitation to surface epitopes and even to antibodies to specific receptors is that the presence of these markers does not equate to functional interaction with the extracellular matrix, with the latter having direct impact on cell phenotype. In support of this, there is consensus in the Glioblastoma field that the tumor microenvironment is central to dictating the cellular state. Given the failure to effectively treat Glioblastoma, and other tumors, based on prior characterization strategies, it is believed that there is an opportunity for a shift in how tumor cells are characterized. Herein, the feasibility is demonstrated and the use of a set of nanoparticle- peptide probes that will differentiate tumor subpopulations by extracellular matrix-binding affinity, rather than traditional, clinically irrelevant cell surface biomarkers, is validated. Develop and characterize nanoparticle-peptide probes based on binding to extracellular matrix peptide motifs. Data has demonstrated that peptides with specificity for extracellular matrix receptors in living cells have been identified based on the interaction of peptide-functionalized nanoparticles (Sloan-Dennison S et al. Chem Sci. 2019, 10(6),1807-15; Sloan-Dennison S et al. Analyst. 2019, 144(18), 5538-46). A toolbox of probes can be developed for cell population sorting by functionalizing the nanoparticles with the amino acid epitopes characteristic of the different extracellular matrix components (fibronectin, laminin, collagens), guided by experience in extracellular matrix cancer interactions through custom extracellular matrix-supported tumor organoid work. Prior results demonstrated that different receptors binding a similar epitope can be differentiated by biophysical approaches (Xiao L et al. Analytical chemistry, 2016, 88(12), 6547-53). The combination of protein affinity assays, Raman imaging, and particle tracking experiments can enable a statistically significant classification of each class of binding interaction found in tumor cell populations. This set of experiments can develop a set of probes with > 90% selectivity for the receptors that bind extracellular matrix components fibronectin, laminin, and collagen IV. Demonstrate cell sorting based on correlation between extracellular matrix binding and functional cell phenotype. The extracellular matrix epitopes will be attached to magnetic beads for cell sorting. Cells sorted using extracellular matrix binding motifs will be derived from established patient-derived organoid models that retain the heterogeneity of the originating tumor tissue. Chemotherapy response, invasion, and cancer stem cell assays are proposed as phenotypic outputs that correlate to poor prognosis. It will also be determined via RNA-seq whether or not Raman imaging data based on peptide-extracellular matrix interactions cluster with a previously reported subpopulation, or by some new distribution. This set of experiments can demonstrate sorting into distinct cell subtypes with divergence of the aforementioned functional assays based on extracellular matrix binding. Current classification technology has not improved survival rates in aggressive and invasive disease. These experiments will validate an initial panel of peptide-nanoparticle probes and establish a workflow that can be used to rapidly identify and sort cells from primary tumors based on functional extracellular matrix interactions. These data will provide rationale to: 1) expand the development of peptide-nanoparticles to other extracellular matrix components, 2) test this methodology with preclinical and clinical samples, and 3) further develop a roadmap for designing combinatorial therapies that target each subpopulation for more comprehensive tumor cell eradication. Example 2 The extracellular matrix drives intratumoral heterogeneity and poor clinical outcomes. The heterogenous nature of many tumors, especially malignant tumors, has long been appreciated histologically and more recent transcriptomic studies have greatly extended this view. This heterogeneity is representative of different cellular subpopulations each with different biological behaviors and differential responses to therapeutic intervention. Hence, any given subpopulation may not effectively respond to a given treatment and therefore drive recurrence and poor clinical outcomes. Additionally, a different subpopulation in that same tumor may harbor a more invasive phenotype and evade surgical resection and/or contribute to metastasis. Central to dictating cellular phenotypes are the interactions of tumor cells with signals within the tumor microenvironment (TME), such as interactions with the extracellular matrix (ECM). A scientific premise of this project is based on the fact that it is known that the physical parameters of the tumor microenvironment – the extracellular matrix – can influence tumor growth and drug responsiveness (Devarasetty M et al. Biofabrication, 2017, 9(2), 021002; Devarasetty M et al. Tissue Eng Part A, 2017, 23(19-20),1026-1041), and can play a role in influencing, facilitating, or even driving tumor progression and subpopulation distribution. Specifically, collagen alignment and composition, fibronectin and laminin concentrations, and other extracellular matrix components, have been documented as playing roles (Miyamoto H et al. Pancreas, 2004, 28(1), 38-44; Neri P et al. Curr Cancer Drug Targets, 2012, 12(7), 776-96; Qin Y et al. Seminars in cancer biology, 2017, 45, 3-12). However, there remains a major lack in technical approaches to isolate these subpopulations in a viable manner based on extracellular matrix interactions as a means to advance the understanding of each subpopulation and hence propose and test treatment paradigms to inclusively target all subpopulations. There are limitations to current live cell sorting methodologies. Purification strategies that capture functional cellular states can allow for a better understanding of the link between intratumoral heterogeneity and malignant cellular phenotypes as well as the response to any current or treatment paradigms. Antibody-based sorting of cell surface markers on live cells is the current standard in the field and has undeniably provided a methodology for the isolation of different populations of cells. This is especially true in the cancer stem cell field whereby this approach has allowed for the functional validation of tumor initiation by cancer stem cells in orthotopic xenograft models over the non-cancer stem cell population from the same tumor (Bao S et al. Nature, 2006, 444(7120), 756-60). However, surface epitopes have also created great controversy in the cancer stem cell field as some groups have demonstrated tumor initiation for both marker positive and negative populations (Shmelkov SV et al. The Journal of clinical investigation, 2008, 118(6), 2111-20; Galdieri L et al. JCI Insight, 2021, 6(4), e128456). One potential limitation to surface epitopes, and even to antibodies to specific receptors, is that the presence of these markers does not necessarily equate to directly dictating a cellular state. Conversely, invasive and aggressive tumors – those with low prognoses – must interact with the extracellular matrix. The cell surface receptors that bind to the extracellular matrix can interact with specific amino acid epitopes (cell adhesion peptides) representative of the various extracellular matrix components: RGD for fibronectin, IKVAV and YIGSR for laminin, GFOGER and xNYYSNS for collagen IV, as well others (Huettner N et al. Trends in Biotechnology, 2018, 36(4), 372-83; Knight CG et al. J Biol Chem. 2000, 275(1), 35-40; Floquet N et al. J Biol Chem. 2004, 279(3), 2091-100). This presents the opportunity to leverage the interactions between tumor cells and their physical tumor microenvironment for the isolation and characterization of tumor cell populations. Glioblastoma is a lethal and heterogeneous disease. Glioblastoma (GBM) is an aggressive, terminal cancer in the brain. Even with maximally aggressive surgery and chemoradiotherapy, median survival for patients with Glioblastoma is 14.5 months (Stupp R et al. The New England journal of medicine, 2005, 352(10), 987-96). These tumors infiltrate normal brain, are surgically incurable, and almost universally recur. Upon recurrence, response rates to standard treatment is less than 5%, leading to a median survival of 8 months (Taal W et al. Lancet Oncol. 2014, 15(9), 943-53). Despite advances in managing cancers, only 4 new treatments have been FDA-approved for Glioblastoma in the last 3 decades. A challenge in translating successful therapies into the clinic is modeling the genetic, epigenetic, and micro- environmental heterogeneity of Glioblastomas, as well as accounting for the blood brain barrier (BBB) (Rybinski B et al. Oncotarget. 2016, 7(44), 72322-42). Glioblastomas evolve spontaneously and in response to treatment, making selection of patient-specific therapies a challenge (Malkki H. Nat Rev Neurol. 2016, 12(4), 190). Glioblastoma is characterized by significant molecular and spatial intratumoral heterogeneity (Sottoriva A et al. Proc Natl Acad Sci U S A. 2013, 110(10), 4009-14). Numerous bulk and single cell transcriptomic analyses have highlighted expression-based subtypes (Verhaak RG et al. Cancer cell. 2010, 17(1), 98-110; Neftel C et al. Cell. 2019, 178(4), 835-49.e21; Couturier CP et al. Nat Commun. 2020, 11(1), 3406; Patel AP et al. Science, 2014, 344(6190), 1396-401; Sottoriva A et al. Proc Natl Acad Sci U S A. 2013, 110(10), 4009-14). Problematically, these coexisting subpopulations evolve over time both under native conditions and in response to therapies. Even after what is considered “successful” treatment, heterogeneous populations of cells survive, and give rise to recurrence of tumors (Heimberger AB et al. Clin Cancer Res. 2003, 9(11), 4247-54; Swartz AM et al. Immunotherapy. 2014, 6(6), 679-90). Despite countless clinical trials, the concept of molecular subtyping has largely failed to yield any transformative improvement to clinical outcomes. Given the failure to effectively treat Glioblastoma, and other tumors, based on prior characterization strategies, there is an opportunity to shift how tumor cell populations are characterized, and in particular Glioblastoma cell subpopulations. The physical tumor microenvironment of Glioblastoma can drive malignancy. The physical tumor microenvironment can play a role in tumor progression and survival but has been overlooked in comparison to consideration of the cellular tumor microenvironment. The physical tumor microenvironment – mechanical matrix properties, but perhaps more importantly, ratios of extracellular matrix components such as hyaluronic acid, fibronectin, laminin, and collagens – has been shown to vary with Glioblastoma region, such as the core, the leading edge, and contrast enhancing versus non-enhancing regions. As the cells occupying these regions, driving the leading edge, and influencing blood brain barrier permeability also exhibit different functions, it stands to reason that there is a correlation between cell subpopulation and functional output. It is hypothesized that specific extracellular matrix binding site-Glioblastoma cell interactions can serve as biomarkers to leverage to categorize Glioblastoma subpopulations. Because the extracellular matrix receptors are on the surface of the cell, these receptors can be readily tagged with molecular probes for identification and classification. A probe set can be developed, based on integrin-binding extracellular matrix peptide sequences covalently bound to magnetic nanoparticles that are amenable to surface enhanced Raman spectroscopy (SERS) and magnetic cellular sorting, thus leveraging the interactions between Glioblastoma cells and their physical brain tumor microenvironment, for the characterization and sorting of Glioblastoma tumor cell populations. Subsequent cellular subpopulations will then be assessed functionally to determine if extracellular matrix-based sorting results in subpopulations with district or overlapping phenotypic behaviors of invasion, drug resistance, stemness, and gene expression. Glioblastoma as a model for validation of the initial probe set will lay the foundation for expansion of the probe set as well as exploration of this methodology in preclinical and clinical samples as well as other tumor types. Given the failure to effectively treat Glioblastoma (and other tumors) based on prior characterization strategies, it is believed that there is an opportunity for a complete shift in how tumor cells are characterized. Rather than traditional surface markers (e.g. EGFR, PDGFRA, CD133, etc.), herein it is proposed to use interactions between Glioblastoma cells and their physical brain tumor microenvironment to evaluate a subpopulation generation workflow. Herein, integrin-binding affinity is explored, but further work can expand to include probes for cell-cell interactions, such as cadherins and tight junction proteins, as well as non-adhesion protein extracellular matrix components that cells physically interact with, such as glycosaminoglycans (e.g., hyaluronic acid). The probe set can enable classification based on the cell integrin binding the extracellular matrix peptide epitope. Recent work showing the SERS signal arises from the receptor binding the nanoparticle will be leveraged to classify probe specificity (Sloan-Dennison S et al. Chemical Science. 2019, 10(6), 1807-15; de Albuquerque CDL et al. Anal Chem. 2020, 92(13), 9389-98; Xiao L et al. Anal Chem. 2017, 89(17), 9091-9; Kim J-Y et al. Anal Chem. 2017, 89(24), 13074-81; Xiao L et al. Anal Chem. 2016, 88(12), 6547-53; Wang H et al. Faraday Discuss.2015, 178, 221-35; Wang H et al. ChemPhysChem. 2014, 15(18), 3944-9; Wang H et al. Analyst. 2013, 138(11), 3150-7). This strategy suggests new technology to characterize surface receptors for subtype classification. The use of peptides to target receptors for identification by SERS has not been widely investigated, suggesting untapped possibilities for a new approach. The extracellular matrix peptide epitopes will be combined with gold-coated iron nanoparticles to enable 1) SERS-based quantification of the distribution of extracellular matrix receptors of cell populations, and 2) magnetic sorting of cells into previously unrealized subpopulations based on their preferential extracellular matrix interactions. Plasmonic magnetic nanoparticles exhibit the SERS effect (Donnelly T et al. Chem Commun. 2014, 50(85), 12907- 10; Wang J et al. ACS Applied Materials & Interfaces, 2016, 8(31), 19958-67; Quaresma P et al. RSC Advances, 2014, 4(8), 3659-67); however, these particles have not been previously used to identify cell surface receptors, much less sort cells based on integrin expression and relative extracellular matrix affinity. This probe set was engineered to not only allow for quantitative characterization through SERS, but to be out-of-the box compatible with commonly used magnetic sorting protocols, enabling a wide pool of end users/customers if this platform is successful. The probe set will provide a new cell subpopulation characterization methodology, based on physical interactions with the extracellular matrix, rather than previous molecular markers, which have failed to yield improved clinical outcomes. Given that tumor cells can interact with extracellular matrix during tumor progression, invasion, and metastasis, there can be higher selectivity for cells associated with malignancy. Bolstering the points of this work are the use of a patient derived Glioblastoma organoid biobank providing cells for the proposed studies, and an underlying extracellular matrix hydrogel platform supporting these organoids that preserves the genomic profiles of the original tumor better than more traditional culture methods. Additionally, retention of expression of integrins was confirmed, while 2D cell culture causes a loss of integrin expression (Figure 1). Approach. Herein, an approach to characterizing Glioblastoma based on interaction with the extracellular matrix through a set of nanoparticle-peptide probes is proposed. Develop and characterize nanoparticle-peptide probes based on binding to extracellular matrix peptide motifs. RGD-Integrin interactions is one of the most utilized extracellular matrix binding interactions studied (Sloan-Dennison S et al. Analyst. 2019, 144(18), 5538-46; Sloan-Dennison S et al. Chemical Science. 2019, 10(6), 1807-15; de Albuquerque CDL et al. Anal Chem. 2020, 92(13), 9389-98; Xiao L et al. Anal Chem. 2017, 89(17), 9091-9; Xiao L et al. Anal Chem. 2016, 88(12), 6547-53; Wang H et al. ChemPhysChem. 2014, 15(18), 3944-9; Xiao L et al. Anal Chem. 2017, 89(17), 9091-9; Huettner N et al. Trends in Biotechnology, 2018, 36(4), 372-83). In this set of experiments, experience with RGD is leveraged to develop nanoparticle (NP) probes containing different cell adhesion peptides for distinct extracellular matrix receptors on cells. Table 1 lists known cell adhesion peptides, the extracellular matrix component they mimic, and their known receptor. As evident in Table 1, integrin receptors are the largest class of proteins found that bind to extracellular matrix components (Reinmuth N et al. Cancer research, 2003, 63(9), 2079-87; Mas-Moruno C et al. Anti-cancer agents in medicinal chemistry, 2010, 10(10), 753-68; Desgrosellier JS et al. Nature reviews Cancer, 2010, 10(1), 9- 22; Cheng S et al. Journal of medicinal chemistry, 1994, 37(1), 1-8; Goswami S. Advances in Biological Chemistry. 2013, 3(2), 224-52; Marelli UK et al. Frontiers in oncology, 2013, 3, 222; Frith JE et al. Stem cells and development, 2012, 21(13), 2442-56). These membrane-spanning proteins trigger signaling pathways that activate proteases and other enzymes that a cell secretes to regulate the extracellular matrix (Shattil SJ et al. Blood. 2004, 104(6), 1606-15; Takagi J et al. Immunological Reviews, 2002, 186, 141-63; van der Flier A et al. Cell and tissue research. 2001, 305(3), 285-98). By focusing on cell adhesion peptides that are derived from specific extracellular matrix components and that target known receptors on cells, probes that screen for extracellular matrix interactions and that can also be correlated with genetic information from the marked cells can be established. Screening for binding motifs offers advantages over antibody-based labelling, by incorporating the protein function into the selection rather than merely the expression. Table 1. Extracellular matrix components and known peptide mimics are listed with their receptors. Cell adhesion peptides can be purchased from commercial vendors (e.g. – Anaspec Inc.) and verified by mass spectrometry upon receipt. Magnetic gold nanoparticles are used (e.g., those that are commercially available from Nanopartz, Inc or CD-Bioparticles). Cell adhesion peptides selected in Table 1 can be functionalized onto gold nanoparticles (AuNPs), providing an improved mimic of the naturally occurring extracellular matrix interface. The diameter of extracellular matrix fibrils are comparable to the dimensions of the gold nanoparticles (Bancelin S et al. Nat Commun. 2014, 5, 4920; Früh SM et al. Nat Commun. 2015, 6, 7275; Erickson HP et al. The Journal of cell biology. 1981, 91(3), 673-8), and it has previously been shown that the avidity effects associated with multiple ligands on a spherical nanoparticles increase the affinity of the particles for protein receptors (Kim J-Y et al. Anal Chem. 2017, 89(24), 13074-81). The affinity and selectivity of the extracellular matrix-mimic nanoparticles is assessed using a suite of biophysical techniques. Figure 2A – Figure 2F: Functionalizing magnetic gold nanoparticles with cell adhesion peptides targets extracellular matrix binding receptors on cells. Direct classical least squares is used to determine a numerical score for binding to a known receptor based on the Raman spectrum. Differences in the spectral correlation score provides a chemically informed comparison of targeting selectivity, providing chemical insight and improved quantification into the targeting of different extracellular matrix peptides. Here, Au-RGDFC is shown to bind more HIILFLHQWO\^WR^Įvȕ3 integrin than bare gold nanoparticle with a p value = 0.008. Additional biophysical assays (e.g., Bradford assay and zeta potential) can be used to determine probe affinity and selectivity. The gold nanoparticles provide improved contrast and SERS signals that can be detected in far-field Raman imaging (Figure 2A). It is known that short amino acid sequences, such as arginine-glycine-aspartic acid (RGD), are ligands for several integrin receptors (Ruoslahti E. Annual review of cell and developmental biology.1996, 12, 697-715) LQFOXGLQJ^Į9ȕ^^ Į,,Eȕ^^^ DQG^Į^ȕ^^^,W^KDV^EHHQ^IXUWKHU^VKRZQ^WKDW^YDULDWLRQV^LQ^WKH ^F-terminus side of the RGD sequence have pronounced effects on receptor recognition (Koivunen E et al. The Journal of cell biology. 1994, 124(3), 373-80). A tryptophan or glycine in this pRVLWLRQ^WDUJHWV^Į^ȕ^^^ZKLOH^D^VHULQH^RU^ DODQLQH^ORFDWHG^KHUH^VKRZV^DIILQLW\^IRU^Į9ȕ^^DQG^Į9ȕ^^^D QG^W\URVLQH^DQG^DUJLQLQH^VXEVWLWXWLRQ^LV^ VSHFLILF^IRU^Į,,Eȕ^ (Cheng S et al. Journal of medicinal chemistry. 1994, 37(1), 1-8). The Raman signature of these different integrins can be determined from in situ experiments with the commercially available purified protein receptors. Previous work enabled discrimination of RGDfC modified gold nanoparticle ELQGLQJ^RI^Į9ȕ^^IURP^Į^ȕ^ (Xiao L et al. Anal Chem. 2016, 88(12), 6547-53). The unique spectral signature of each receptor can be modeled using multivariate statistical approaches (e.g. MCR) (Sloan-Dennison S et al. Chemical Science. 2019, 10(6), 1807-15), Figure 2B, supporting classification based on unique receptor signals. An example of the Raman based specificity assessment is shown in Figure 2D, where cells incubated with RGDFC-nanoparticleV^VKRZHG^D^VWDWLVWLFDOO\^KLJKHU^DIILQLW\^W R^WKH^Į9ȕ^^,QWHJULQ^ receptor than unfunctionalized nanoparticles. The statistical variation observed in the Raman maps will provide an indicator of ligand selectivity. A highly conserved spectrum, such as was SUHYLRXVO\^REVHUYHG^LQ^H[SHULPHQWV^ZLWK^F\FOLF^5*')&^SHS WLGHV^IRU^Į9ȕ^^LQWHJULQ^^ZLOO^VKRZ^ little statistical variance. A promiscuous ligand, such as CisoDGRC, shows a high degree of spectral variance which indicates multiple binding interactions (Xiao L et al. Probing membrane receptors with enhanced Raman imaging. SPIE Nanoscience + Engineering: Nanoimaging and Nanospectroscopy VI; 2018: International Society for Optics and Photonics). The Raman classification will be analyzed to determine p-values for receptor specificity. Binding affinity for free protein and in cells will be determined by complementary techniques. For free protein, decreasing concentrations purified receptors can be incubated with the magnetic nanoparticles and changes in unbound protein concentration can be assessed using absorbance at 280 nm, Bradford assay, or using Krypton™ stain and fluorescence detection (Noble JE. Chapter Two – Quantification of Protein Concentration Using UV Absorbance and Coomassie Dyes. In: Lorsch J, editor. Method Enzymol: Academic Press; 2014. P. 17-26; Ernst O et al. JoVE. 2010, 38, e1918). Changes in zeta-potential with changing protein concentration will further be used to validate binding to the nanoparticles. The binding to purified proteins will be compared with in vitro single particle tracking experiments to assess differences observed from cells (Xiao L et al. Anal Chem. 2017, 89(17), 9091-9). The affinity of the extracellular matrix-mimic nanoparticles will be compared with traditional antibody functionalized magnetic gold nanoparticles on cells as a control. Preliminary data suggests that the peptide-functionalized nanoparticle probes bind extracellular matrix receptors in cells (Sloan-Dennison S et al. Analyst. 2019, 144(18), 5538-46). The corona inhibition associated with RGD will not persist with other ligands. The inhibition is typically associated with a zwitterionic surface and has been observed for other amino acid coatings. Multiple options will be investigated, noting new corona prevention technologies is also an important advance. Specific receptors cannot be identified by Raman/SERS. While the absolute identity of the binding site cannot be determined from the SERS spectrum, the unique composition and configuration of the amino acids gives rise to a unique fingerprint (Nguyen Anh H et al. Reviews in Analytical Chemistry. 2017, 36(4), 20160037). Thus clustering and multivariate analysis of the SERS signal will enable determination of a specific or variable binding interactions that can be correlated to off-target effects. If larger SERS enhancements are needed, gold nanostars (AuNS) will be synthesized using a seed mediated synthesis and functionalized as previously reported (Sloan-Dennison S et al. Chemical Science. 2019, 10(6), 1807-15). Other extracellular matrix components and receptors may drive classification. Tenascin C and hyaluronic acid are known to be important as well, though these components utilize glycans for recognition (Lam D et al. Sci Rep-Uk. 2019, 9(1), 4159). Initially, peptide mimics of the three major components have been focused on, but glycan recognition can also be explored. This set of experiments will develop extracellular matrix peptide modified nanoparticle probes to assess binding selectivity to cells. It is expected to measure SERS spectra from nanoparticle probes where each spectrum can be correlated to the binding partner of ligand functionalized to the gold nanoparticle. These experiments can develop a set of probes with > 90% selectivity for receptors that bind the extracellular matrix components Fibronectin, Laminin, and Collagen IV. Demonstrate cell sorting based on correlation between extracellular matrix binding and functional cell phenotype. There is extensive expertise in extracellular matrix biology and designing, engineering, and implementing extracellular matrix hydrogel biomaterials in a variety of tissue engineering applications, such as generation of patient-specific tumor organoids from a wide range of tumor types, including Glioblastoma (Forsythe S et al. Annals of biomedical engineering.2020, 48(3), 940-52; Maloney E et al. Micromachines. 2020, 11(2), 208; Mazzocchi A et al. ACS Biomaterials Science & Engineering. 2019, 5(4), 1937-43; Mazzocchi AR et al. Sci Rep-Uk. 2018, 8(1), 2886; Votanopoulos KI et al. Annals of Surgical Oncology. 2020, 27(6), 1956-67; Votanopoulos KI et al. Annals of Surgical Oncology. 2019, 26(1), 139- 47). In a complementary fashion, there is also extensive expertise in Glioblastoma tumor cell biology, tumor heterogeneity, patient-derived xenograft models, and glioma stem cell cultures and associated biological assays. In this experiment set, an established portfolio of model systems and assays will be leveraged to functionally, phenotypically, and genetically validate the peptide probe toolkit developed above. The extracellular matrix peptide modified nanoparticle probes will be utilized to bin heterogeneous Glioblastoma populations into subpopulations. Functionally, Glioblastoma is defined by its invasive nature, differential drug response – primarily to temozolomide (TMZ) – and presence of glioblastoma stem cells (GSCs). 3D invasion assays will be employed that are largely unidirectional, using an extracellular matrix hydrogel platform that has supported previous patient-derived tumor organoid (PTO) systems (Forsythe S et al. Annals of biomedical engineering.2020, 48, 940-952; Maloney E et al. Micromachines (Basel). 2020, 11(2), 208; Forsythe SD et al. Ann Surg Oncol. 2020, 27(13), 4950-60; Mazzocchi A et al. ACS Biomater Sci Eng. 2019, 5(4), 1937-43; Mazzocchi AR et al. Sci Rep. 2018, 8(1), 2886; Votanopoulos KI et al. Ann Surg Oncol. 2020, 27, 1956-1967; Votanopoulos KI et al. Ann Surg Oncol. 2019, 26(1), 139-47; Votanopoulos KI et al. Ann Surg Oncol. 2020, 27, 1968-1969) (thiolated HA, methacrylated collagen, with synthetically-modified fibronectin (FN) and laminin (LMN)) (Aleman J et al. bioRxiv. 2021, 458584), and which recently has been shown to preserve genomic profiles of the originating tumor (Figure 3). This will be a superior invasion assay compared to more common 2D migration assays and spheroid outgrowth invasion assays. Drug responsiveness will be performed in drug screens with temozolomide, the primary drug administered to Glioblastoma patients. Phenotypically, the relative stemness of each sorted subpopulation will be evaluated. In vitro limiting dilution assays and serial tumorsphere formation assays will be performed, as routinely done (Summers MK et al. Cancer Letters. 2021, 499, 232-42; Dermawan JKT et al. Cancer research. 2016, 76(8), 2432-42; Venere M et al. Science Translational Medicine. 2015, 7(304), 304ra143). These assays are surrogates for the key glioblastoma stem cell phenotype of self-renewal and allow for the quantification of stem- cell frequency (Hu Y et al. Journal of Immunological Methods. 2009, 347(1), 70-8). Lastly, RNAseq will be performed on each of the sorted subpopulations and genomic profile differences will be compared. These assays will determine the extent to which the extracellular matrix affinity-sorted subpopulations are truly unique with clear functional, phenotypic, and genetic differences. To challenge the peptide probe set, Glioblastoma cells originating from human biospecimens will be employed to ensure heterogeneity observed in vivo. These cells have been maintained as Glioblastoma patient-derived tumor organoids and biobanked. To date, over 30 sample sets have been biobanked. As described above and in Figure 3, the extracellular matrix hydrogel 3D culture approach preserves heterogeneity and subpopulation distribution of the originating tumor. As such, there is a ready to use source of heterogeneous Glioblastoma cell populations for these studies. The use of iron core nanoparticles in the peptide probe set enables SERS-based quantification of subpopulations with extracellular matrix-binding proclivities, and subsequent sorting using standard magnetic cell sorting columns. Patient-derived cells from the Glioblastoma organoid biobank will be sorted hierarchically by pulling the predominantly laminin-binding cells with laminin peptide probes, then predominantly fibronectin-binding cells, and predominantly collagen-binding cells in the same manner. These three subpopulations will be analyzed using an invasion assay, temozolomide drug screen, stemness assay, and by RNAseq. As part of this work, the hierarchical sorting order will be alternated to evaluate overlap between subpopulations, and potentially which sorting order results in the most distinction between sorted subpopulations. 3D invasion assay. The current invasion assay is a modification of a previously published assay, which places a tumor cell-hydrogel foci within a cell free hydrogel, enabling tracking of fluorescently labeled cells outwards (Skardal A et al. Biotechnol Bioeng.2016, 113(9), 2020-32). A similar method will now be used, based on in situ photopatterning multi-region tumor constructs within microfluidic devices (Rajan SAP et al. Advanced Biosystems. 2020, 4(4), 1900273), but advancing this technique to increase throughput and automation by using digital light processing (DLP) bioprinting. Two hydrogel solutions – one cell and label free, and the other labeled by an AlexaFluor 488 maleimide that links to the thiol groups of the hyaluronic acid component of the hydrogel in which tumor cells labeled with a cell membrane intercalating dye (DiD, ThermoFisher) are suspended and printed simultaneously to create adjacent, seamlessly crosslinked adjacent regions. The AlexaFluor bound to one of the hydrogel regions creates a demarcation from which invasion distances of invading cells can be measured to quantitatively assess invasion kinetics by both average distances migrated and velocities (Figure 4A-Figure 4B). Imaging will be performed using an EVOS M5000 for global assessment of bulk migration, and a resonant scanning confocal microscope will provide detailed imaging. Temozolomide drug screen. Sorted cell population-based organoids will be bioprinted into 96-well plates in the HA-collagen-LMN-FN extracellular matrix hydrogel, as described elsewhere (Maloney E et al. Micromachines (Basel). 2020, 11(2), 208; Clark CC et al. Bioprinting. 2019, 16, e00058; Mazzocchi A et al. Applied Physics Reviews. 2019, 6, 011302). Organoids will be treated with 0, 10, 100, 1000, or 10000 ^M temozolomide (TMZ). ATP quantification (proportional to cell number) and LIVE/DEAD imaging will be performed on days 1, 4, and 7 after treatment, with media/drug changes on day 4. This experiment is the same as those that have been performed with temozolomide in Glioblastoma patient-derived tumor organoids (Maloney E et al. Micromachines (Basel). 2020, 11(2), 208) or other drug compounds in patient-derived tumor organoids of other tumor types (Forsythe S et al. Annals of biomedical engineering.2020, 48, 940-952; Maloney E et al. Micromachines (Basel). 2020, 11(2), 208; Forsythe SD et al. Ann Surg Oncol. 2020, 27(13), 4950-60; Mazzocchi A et al. ACS Biomater Sci Eng. 2019, 5(4), 1937-43; Mazzocchi AR et al. Sci Rep. 2018, 8(1), 2886; Votanopoulos KI et al. Ann Surg Oncol. 2020, 27, 1956-1967; Votanopoulos KI et al. Ann Surg Oncol. 2019, 26(1), 139-47; Votanopoulos KI et al. Ann Surg Oncol. 2020, 27, 1968-1969). In preliminary studies using the RGDFC-nanoparticles (fibronectin affinity) for cell sorting of patient-derived glioblastoma cells, a limited scope temozolomide drug screen was performed that showed a divergence in temozolomide drug response between patient-derived tumor organoids formed with cells with high FN affinity and cells with low FN affinity (Figure 4C). This potential confirmation that the extracellular matrix nanoparticle sorting strategy can identify true tumor subpopulations that are functionally distinct. Stemness. Limiting dilution assays will be performed by serial dilution of gold nanoparticle-sorted Glioblastoma subpopulations to seed cells in suspension in low-attachment plates at 20, 10, 5, or one cell per well followed by scoring for sphere formation 10-14 days later. Stem-cell frequency will be calculated using a freely available webtool called ELDA (extreme limiting dilution analysis) (Hu Y et al. Journal of Immunological Methods.2009, 347(1), 70-8). For serial tumorsphere formation to evaluate continued self-renewal capabilities central to the glioblastoma stem cell state, gold nanoparticle-sorted Glioblastoma subpopulations will be seeded in low attachment plates at one cell per well and allowed to form tumorspheres. Resulting spheres will be dissociated to a single cell suspension and again seeded at one cell per well and challenged to form secondary tumorspheres. This process will be performed for a total of 4 rounds of tumorsphere formation with quantification of the percentage of cells that were tumorsphere forming determined for each serial sphere formation. RNAseq. Following sorting, subpopulations will be evaluated by RNAseq to determine if extracellular matrix interaction-based sorting generates genetically distinct subpopulations. Data sets will be assessed via cluster analysis to quantitatively distinguish differences between each population. RNA will be isolated (Qiagen Rneasy) from each extracellular matrix interaction- sorted subpopulation and RNAseq will be performed for the 3 populations in question and analyzed. Notably, in preliminary data in which glioblastoma subpopulations that were isolated with fibronectin-mimicking probes versus the remaining cell populations were analyzed, qRT- PCR shows a clear divergence in gene expression across a number of genes often used to characterize glioblastoma stem cells and glioblastoma (Figure 4D), further corroborating the data in Figure 4C suggesting that extracellular matrix component affinity may be a new and effective way at characterizing and isolating distinct glioblastoma subpopulations. Preliminary data suggests the organoid cultures maintain the extensive heterogeneity seen in primary patient samples, so the diversity of receptors is expected to be expressed and interact with the different extracellular matrix nanoparticles. The current invasion assay utilizes the HA, collagen, fibronectin, and laminin hydrogel, and thus should support quantitative motility measurements of any cell population. However, this may not be sufficiently precise. As an alternative, HA hydrogels can be prepared with i) collagen, ii) fibronectin, or iii) laminin, thus challenging nanoparticle-sorted subpopulations with only one extracellular matrix protein with which to interact for invasion. This should force deviation of quantitative motility kinetics between subpopulations. Additionally, a chemokine can be added to the cell free hydrogel if it is desired to accelerate invasion. Notably, the HA component of the hydrogel is already modified with heparin pendant chains, allowing immobilization of heparin-binding growth factors/chemokines, as has been demonstrated extensively (Murphy SV et al. Stem Cells Transl Med. 2020, 9(1), 80-92; Murphy SV et al. Stem Cells Transl Med. 2017, 6(11), 2020-32; Skardal A et al. Bioprinting. 2018, 10; e00026; Skardal A et al. Acta Biomater. 2015, 25, 24-34; Skardal A et al. J Biomed Mater Res B Appl Biomater. 2017, 105(7), 1986-2000; Skardal A et al. Scientific Reports. 2017, 7, 8837; Skardal A et al. Biomaterials. 2012, 33(18), 4565-75). The stem cell assays employed do not test for in vivo tumor initiation. Although in vivo limiting dilution assays in immunocompromised mice can be used to query for the stemness property of tumor initiation, the in vitro assays proposed are well established and accepted in the cancer stem cell field to test for the stem cell state without the need of incorporating mouse models into the proposed studies at this time. The RNA-seq results for the nanoparticle-sorted Glioblastoma subpopulations may not yield distinct transcriptional signatures. The receptors associated with the different extracellular matrix components are each linked with distinct downstream pathways so differential transcriptional signatures should be obtained but overlap is possible, and this will be elucidated via analysis of the sequencing results. This set of experiments can correlate cellular phenotypes to the subpopulations sorted by extracellular matrix profiles. These experiments focus on the invasion, drug response, stemness phenotypes, and genetic profiles. Successful nanoparticle-based sorting to real subpopulations will be achieved if within each assay type, the collagen-, fibronectin-, and laminin-specific subpopulations display statistically distinct (p<0.05) Gaussian distributions of quantitative invasion kinetics, drug response, tumorsphere scores, and the 1,000 most variably expressed genes. This work will validate a panel of peptide-nanoparticle probes and establish a workflow that can be used to rapidly identify and sort cells from primary tumors based on functional extracellular matrix interactions. These data will provide rationale to: 1) expand the development of peptide-nanoparticles to other extracellular matrix components, 2) test this methodology with preclinical and clinical samples, and 3) further develop a roadmap for designing combinatorial therapies that target each subpopulation for more comprehensive tumor cell eradication. Given that molecular subtyping and newer stemness- or immune profiling-based characterization of glioblastoma subpopulations have largely failed to translate into clinical advancement, this technology offers a new approach. Moreover, it is easily translatable to other cancer types, such as in additional validation studies using colorectal cancer cell subpopulations being tested. Example 3 – Classification of Glioblastoma Cancer Stem Cells Using Magnetically Sorted Surfaced Enhanced Raman Spectroscopy and Extracellular Matrix Peptide Mimics Described herein are experiments related to glioblastoma cellular subtyping and sorting using extracellular matrix peptide mimics and surface enhanced Raman spectroscopy. Abstract. Glioblastoma is an extremely heterogeneous and aggressive brain cancer with low survivability which has been found to have cellular subtypes that mimic various stages of brain development. Treating glioblastoma is extremely difficult because within the heterogeneous nature of the cancer, the prevalence of cancer cells with stem cell like properties which self-propagate, and readily adapt to their environment, has led to cancer recurrence in many patients after treatment. The current method for sorting and classifying these cells uses magnetic nanoparticles functionalized with various antibodies such as CD133 and CD 44 to sort via magnetically activated cell sorting and use of RNA sequencing to determine genetic information. This sorting method has been shown to be unable to adequately separate and classify these cellular subtypes. The tumor microenvironment offers a potential solution to this problem. Fibronectin, laminin, and collagen are all important extracellular matrix (ECM) components that interact with glioblastoma cells through heterodimeric cell surface receptors known as integrins. Functionalizing gold coated magnetic nanoparticles with small peptide mimics for these extracellular matrix components allows for a robust extracellular matrix based sort and access to Surface enhanced Raman spectroscopy (SERS). SERS is a nondestructive ultrasensitive technique which offers a direct method for the identification of these cellular subtypes based on their observed SERS signal. Using SERS, the functionalization of various small peptide extracellular matrix mimics for the three main extracellular matrix components, such as the well-studied Arg-Gly-Asp (RGD) peptide for fibronectin mimic, can be confirmed and their interaction with their corresponding integrin can be observed allowing for the creation of a library of vibrational spectra that can be used in cell sorting. These peptide-functionalized plasmonic magnetic nanoparticles can then be incubated with glioblastoma cells and serially sorted as to remove the cellular subtypes with a high prevalence of the targeted integrin one at a time to create sorted populations. The sorted cells can then be collected and mapped using SERS with little prep. The resulting maps allow for the classification of different cellular subtypes based on the presence of the integrin signals seen throughout the cell. Introduction. Glioblastoma is the most common primary, malignant brain tumor (Bae K et al. Anal. Chem. 2021, 93(4), 2377–2384). Currently, there are 4 accepted cellular subtypes (Verhaak RGW et al. Cancer Cell 2010, 17(1), 98–110): mesenchymal, neural, proneural, and classical. However, the current cellular subtype classification is ineffective and has made treatment difficult. Cancer cells interact with extracellular matrix (ECM) components in the brain. Cells interact through unique surface receptors. Extracellular matrix components (e.g., fibronectin, laminin, and collagen) can be mimicked with small peptides used for sorting. Methods. Surface Enhanced Raman Spectroscopy was utilized for binding identification. Magnetic nanoparticles were utilized for cellular sorting. An example particle is shown in Figure 5. An iron core provides paramagnetic activity and generation of hot spots. A gold shell provides plasmonic effects (Figure 6) for SERS. Cellular Sorting Methods. Peptide functionalized magnetic nanoparticles were incubated with cells. Cells expressing desired cellular receptors were magnetically sorted out using magnetic activated cellular sorting (MACS) (Figure 7; positive selection (yellow)). Cells without desired receptor are not separated and flow through (Figure 7; Negative selection (purple)). Fibronectin Mimic Characterization. Results are shown in Figure 8 – Figure 10. The SERS results show a peak at 1001 cm ^-1 for the peptide mimic (Figure 8). Zeta Potential (Figure 9) shows a positive increase as the peptide mimic is protonated. Target protein signal corresponds to nearby amino acids (Stiles PL et al. Annual Review of Analytical Chemistry 2008, 1(1), 601–626; Xiao L et al. Anal. Chem. 2016, 88(12), 6547–6553) (Figure 10). Laminin Mimic Characterization. Results are shown in Figure 11 – Figure 13. There was no discernable SERS signal for laminin peptide mimic, but there was a reproducible and unique spectral response for the target receptor (Figure 11 and Figure 12). A unique protein binding signal was observed after incubation with a target protein (Figure 13). Cellular Sorting using extracellular matrix Mimics. Fibronectin protein binding pocket was determined using Multivariate Curve Resolution (MCR) (Figure 14). Cells were sorted using extracellular matrix mimic functionalized magnetic nanoparticles: Positive Selection – Sorted for target protein; Negative Selection – Unsorted Cells. Positively sorted cells show significantly more binding of nanoparticles than negatively sorted cells (Figure 15 and Figure 16). Underlying protein spectra are present in high scoring spots. There were problems with burning from heat at areas with nanoparticles. Conclusions. Extracellular matrix peptide mimic functionalized magnetic nanoparticles were designed. Sorting on cells shows a difference between selected cells and non sorted cells. Other goals include serial sorting for multiple peptide mimics and RNA sequencing on sorted populations. Example 4 Described herein are magnetic nanoparticles and methods of use thereof. The Magnetic nanoparticles have a magnetic core typically iron in composition with a gold shell for the plasmonic properties and then a layer of peptide functionalized onto to them (Figure 5). Cyclic RGDfC Functionalization of Plasmonic Magnetic Nanoparticles – Fibronectin Mimic: SERS spectra taken of nonfunctionalized (bare) and Cyclic RGDfC functionalized gold coated magnetic nanoparticles (Figure 17) (633 nm, 1s, 0.14 mW, 50x, pinhole). Appearance of 1001 cm ^-1 peak in Figure 17 corresponds to phenylalanine. Zeta potential measurements taken of nonfunctionalized (bare) and Cyclic RGDfC functionalized gold coated magnetic nanoparticles at pH 4 and pH 7 which is above and below the isoelectric point of the peptide (Figure 18). Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles (Figure 19 and Figure 20). As seen in Figure 20 a small plasmon shift (575 nm to 577 nm) is observed once peptide is bound to magnetic nanoparticles. Cyclic RGDfC (Fibronectin Mimic) – Protein Binding – Fibronectin Mimic: SERS spectra taken of Cyclic RGDfC functionalized gold coated magnetic nanoparticles incubated ZLWK^,QWHJULQ^Įvȕ3 and the corresponding SERS peak assignments (Figure 21). Bradford assay DEVRUEDQFH^VSHFWUD^RI^VWRFN^,QWHJULQ^Į _v ȕ ₃ (blue), Cyclic RGDfC Functionalized magnetic nanoparticles (red), Supernatant after magnetic separation (yellow), Cyclic RGDfC with Integrin Įvȕ3 incubation remaining after magnetic separation (Figure 22). ,QWHJULQ^Į _v ȕ ₃ Incubation with Bradford Assay^^,QWHJULQ^Į _v ȕ ₃ calibration curves using Bradford protein binding assay with corresponding LOD for the protein (Figure 23 and Figure 24). CDPGYIGSR Functionalization of Plasmonic Nanoparticles – Laminin Mimic: SERS spectra taken of nonfunctionalized (bare) and CDPGYIGSR functionalized gold coated magnetic nanoparticles (Figure 25). Zeta potential measurements taken of nonfunctionalized (bare) and CDPGYIGSR functionalized gold coated magnetic nanoparticles at pH 4 and pH 7 which is above and below the isoelectric point of the peptide (Figure 26). Absorbance measurements of Cyclic RGDfC functionalized and Bare gold coated magnetic nanoparticles (Figure 27 and Figure 28). As seen in Figure 28 a small plasmon shift (575 nm to 577 nm) is observed once peptide is bound to magnetic nanoparticles. CDPGYIGSR Integrin Incubation – Laminin Mimic: SERS spectra taken of Cyclic CDPGYIGSR functionalized gold coated magnetic nanoparticles incubated with InWHJULQ^Į ₄ ȕ ₁ and the corresponding SERS peak assignments (Figure 29). CDPGYIGSR – Integrin Specificity – Laminin Mimic: SERS spectra taken of &'3*<,*65^IXQFWLRQDOL]HG^JROG^FRDWHG^PDJQHWLF^QDQRSDU WLFOHV^LQFXEDWHG^ZLWK^,QWHJULQ^Į ₄ ȕ ₁ DQG^,QWHJULQ^Į ₆ ȕ ₁ (Figure 30) and a table of their corresponding peak assignments (Table 2) (Movasaghi Z et al. Applied Spectroscopy Reviews 2007, 42 (5), 493–541). Table 2. Peak assignments. Cellular Sorting – Using Cyclic RGDfC Fibronectin Mimic. Microscope images of negatively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles are shown in Figure 31 and Figure 32. Microscope images of positively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles are shown in Figure 33 and Figure 34. A microscope image of negatively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles is shown in Figure 35. Microscope images (bright field and dark field) of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 1 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles are shown in Figure 36 and Figure 37. Microscope images (bright field and dark field) of negatively sorted cells with the incubation conditions of 1 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles are shown in Figure 38 and Figure 39. Microscope images (bright field and dark field) of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 5*10 ⁸ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles are shown in Figure 40 and Figure 41. Microscope images (bright field and dark field) of negatively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles are shown in Figure 42 and Figure 43. A microscope image of positively sorted cells with the incubation conditions of 2 times the Miltenyi bead concentration (about 1*10 ⁹ nanoparticles) and 2 hr incubations of RGDfC Functionalized gold coated magnetic nanoparticles is shown in Figure 44. Cellular Sorting. ,QWHJULQ^Į _v ȕ ₃ Multivariate curve resolution component generated from LQFXEDWLQJ^,QWHJULQ^Į _v ȕ ₃ with Cyclic RGDfC functionalized magnetic nanoparticles in situ (Figure 45). Cellular Sorting – Scores Maps. 6FRUHV^RI^,QWHJULQ^Įvȕ3 on SERS cell maps FRUUHVSRQGLQJ^WKH^,QWHJULQ^Į _v ȕ ₃ multivariate curve resolution component for the protein on both a negatively sorted cell and a positively sorted cell (Figure 46 and Figure 47). The positive cell has the presence of higher scoring protein signal within the map. CDPGYIGSR Studies. Results are shown in Figure 48 – Figure 50 (633 nm, 1s, 0.14 mW, 50x, pinhole). Replicated results are shown in Figure 51 – Figure 53. YIGSR Studies. Weak SERS Scattering/ Usually nearly 0 or inconsistent. Tyrosine Absorption around 280 nm. Obvious Change in Zeta Potential showing functionalization. Protein binding via Bradford Studies and SERS suggest no binding right now. CDPGYIGSR Bradford Studies. Results are shown in Figure 54 and Figure 55. Cyclic RGDfC Updates – Cellular Populations used for the Sort. Results are shown in Figure 56 and Figure 57 (633 nm, 1s, 0.14 mW, 50x, pinhole). Cyclic RGDfC Integrin Studies. Results are shown in Figure 58 (633, 1s, 0.5 NA (50X), pinhole, 0.14 mW) and Figure 59. Cyclic RGDfC Integrin AvB3 Binding. Results are shown in Figure 60 and Table 3. Cyclic RGDfC. Cyclic RGDfC Functionalization shown with absorption at the 260 nm region indicative of Phenylalanine as well as the 1003 cm ^-1 scattering peak (Figure 61 and Figure 62). Table 3. Peak assignments. aMovasaghi Z et al. Applied Spectroscopy Reviews 2007, 42 (5), 493–541 bXiao L et al. Anal. Chem. 2016, 88 (12), 6547–6553 Example 5 – Genetic and Function Data Following Extracellular Matrix Affinity Sorting (Fibronectin-Functionalized Particles) qRT-PCR analysis of several key genes important in Glioblastoma (CD44, EGFR, and Olig2) show dramatic differences between expression levels between cell subpopulations selected with FN peptide-functionalized beads and the remaining subpopulations (Figure 63- Figure 64). Analysis of additional genes by qRT-PCR show dramatic differences between expression levels between cell subpopulations selected with FN peptide-functionalized beads and the remaining subpopulations (Figure 65). Following FN-bead sorting, cells were used to create hyaluronic acid and gelatin-based 3D tumor constructs or “organoids” (thiolated hyaluronic and thiolated collagen, crosslinked with polyethylene glycol diacrylate in the presence of a photoinitiator under brief UV light pulses), which were subjected to drug screens with temozolomide (TMZ) (Figure 66). “1 hr 1x NEG” indicates no bead was used, while “1 hr 1x POS” indicates the FN peptide-functionalize beads were used. While both organoid sets did ultimately respond to temozolomide at higher concentrations, at a moderate concentration (100 uM), the FN bead-selected cell subpopulation was significantly more resistant to the drug (statistical significance: * p < 0.5), suggesting true biological differences between these subpopulations. Example 6 Disclosed herein in some examples are extracellular matrix mimic magnetic nanoparticles for tumor classification. The compositions disclosed herein can comprise magnetic nanoparticles that have peptides attached designed to mimic the extracellular matrix found in the brain (or other parts of the body). These particles show affinity for cells derived from brain tumors. The treatment for brain tumors, specifically glioblastomas, have not been largely successful due to the variety of cell types comprising the tumor. These probes select for cell types that will bind to the extracellular matrix, which is a step in the growth and metastasis of the tumor. The peptides on the magnetic particles bind to proteins receptors on the surfaces of cells and then the cells that bind can be selected using a magnetic field. Results suggest these probes can provide an improved classification of the types of cells found in the tumor which can improve treatment and increase survival rates. Data has been obtained including synthesis of 2 separate probes, 1 of which has been demonstrated to successfully sort cells after binding. Glioblastoma derived cells were successfully sorted using a peptide that mimics fibronectin attached to magnetic nanoparticles. Example 7 – Cancer Cell Targeting, Magnetic Sorting, and SERS Detection through Cell Surface Receptors Abstract. Integrins are cellular surface receptors responsible for activation of many cellular pathways in cancer. These integrin proteins can be specifically targeted by small peptide sequences that offer the potential for the differentiation of cellular subpopulations using magnetically assisted cellular sorting techniques. By adding a gold shell to the magnetic nanoparticles, these integrin-peptide interactions can be differentiated by Surface enhanced Raman spectroscopy (SERS) providing a quick reliable method for on target binding (Figure 67). Herein, the ability to differentiate the peptide-protein interactions of the small peptides CDPGYIGSR and Cyclic RGDfC functionalized on gold coated magnetic nanoparticles with the integrins they are known to bind to using their SERS signal is demonstrated. SW 480 and SW 620 colorectal cancer cells known to have the integrins of interest were then magnetically sorted using these functionalized nanoparticles suggesting differentiation between the sorted populations and integrin populations among the 2 cell lines. Introduction. The formation of heterogeneous tumors is believed to have significant impact on treatment efficacy in cancer. As one example, glioblastoma (GBM) is a heterogeneous tumor believed to have at least 4 cellular subtypes, making it difficult to treat and characterize (Verhaak RGW et al. Cancer Cell 2010, 17 (1), 98–110). Likewise, in other cancer types, such as colorectal cancer, subpopulations can be stratified in a number of ways, including by accrual of mutations such as WNT, KRAS and BRAF, which can result in changes in proliferation and invasion potential. In cancer, attachment to, and detachment from, the extracellular matrix (ECM) regulates cellular signaling and invasiveness of the tumor (Hatoum A et al. Cancer Manag Res 2019, 11, 1843–1855). The extracellular matrix includes matrix molecules with which cellular receptor proteins interact to influence cellular signaling pathways, such as migration, maturation, differentiation, survival, and tissue homeostasis (Mohiuddin E et al. Am J Cancer Res 2021, 11 (8), 3742–3754). Since the extracellular matrix can regulate cellular differentiation, studying the extracellular matrix-cell interactions offers a potentially useful tool for differentiating tumor subpopulations allowing for more targeted treatments. Three important and prominent structural components of the extracellular matrix are vitronectin, fibronectin and laminin, which bind cancer cells through integrin proteins and signal the activation of cellular pathways (Hatoum A et al. Cancer Manag Res 2019, 11, 1843–1855; Mohiuddin E et al. Am J Cancer Res 2021, 11 (8), 3742–3754; Oomura A et al. Virchows Arch A Pathol Anat Histopathol 1989, 415 (2), 151–159; Bachman H et al. Adv Wound Care (New Rochelle) 2015, 4 (8), 501–511; Belkin AM et al. Microsc Res Tech 2000, 51 (3), 280–301). Integrins are a family of cellular surface receptors that mediate the attachment of cells to the extracellular matrix and have specificity in their binding based on the combination of Į and ȕ domains (Barczyk M et al. Cell Tissue Res 2010, 339 (1), 269–280). In order to target specific integrins and their interactions with the extracellular matrix, small peptides such as the Arg-Gly- Asp (RGD) and Tyr-Ile-Gly-Ser-Arg (YIGSR) can mimic vitronectin or fibronectin and laminin respectively (Boateng SY et al. American Journal of Physiology-Cell Physiology 2005, 288 (1), C30–C38; Cheng S et al. J. Med. Chem. 1994, 37 (1), 1–8; Massia SP et al. Journal of Biological Chemistry 1993, 268 (11), 8053–8059; Yoshida N et al. Br J Cancer 1999, 80 (12), 1898–1904). The RGD peptide sequence targets two integrins, integrin Įvȕ3 and integrin Į5ȕ1 , that are correlated with disease progression (Danhier F et al. Mol. Pharmaceutics 2012, 9 (11), 2961– 2973; Desgrosellier JS et al. Nat Rev Cancer 2010, 10 (1), 9–22; Mao Y et al. Cell Communication & Adhesion 2006, 13 (5–6), 267–277; Bowditch RD et al. Journal of Biological Chemistry 1994, 269 (14), 10856–10863). Integrin Į _v ȕ ₃ is associated with enhanced cellular motility and migration and resistance to apoptosis (Danhier F et al. Mol. Pharmaceutics 2012, 9 (11), 2961–2973; Sheldrake HM et al. Curr Cancer Drug Targets 2009, 9 (4), 519–540), while integrin Į ₅ ȕ ₁ is important in assembling the fibronectin matrix and initiating the cellular attachment to the extracellular matrix (Koivunen E et al. J Cell Biol 1994, 124 (3), 373–380). The YIGSR sHTXHQFH^RQ^WKH^RWKHU^KDQG^KDV^EHHQ^VKRZQ^WR^ELQG^WR^WKH^ȕ1 integrin domain, VSHFLILFDOO\^WR^WKH^LQWHJULQ^Į ₆ ȕ ₁ EXW^DOVR^KDV^EHHQ^UHSRUWHG^WR^ELQG^WR^LQWHJULQ^Į ₄ ȕ ₁ (Sanchez- Esteban J et al. American Journal of Physiology-Lung Cellular and Molecular Physiology 2006, 290 (2), L343–L350; Maeda T. et al. The Journal of Biochemistry, 1994, 115(2), 182-189; Schaff M et al. Circulation 2013, 128 (5), 541–552). While these small peptide mimics have been used in cell spreading experiments (Massia SP et al. Journal of Biological Chemistry 1993, 268 (11), 8053–8059; Yu C et al. PNAS 2011, 108 (51), 20585–20590) and to provide cell attachment points in 3D hydrogel cultures, their ability to be grafted or functionalized onto nanoparticles make them ideal probes for studying integrin binding to a given cancer cellular population. Raman probes (Huang R et al. Theranostics 2016, 6 (8), 1075–1084) or fluorescent tags (Wu PH et al. IJN 2017, 12, 5069– 5085) are often used in conjunction with cellular imaging to assess a cell’s binding phenotype. These methods provide intense signals, but with the loss of the chemical information at the binding site that is useful for analyzing specificity of the binding and the ability to differentiate the possible interactions. SERS is a nondestructive spectroscopic technique that is insensitive to water, making it useful for studying biological systems (Lai H et al. Journal of Materials Science 2018, 53 (12), 8677–8698). It has been shown previously that the RGD peptide has been functionalized onto gold nano-stars and the unique spectral fingerprint has allowed for label free identification of the proteins (Sloan-Dennison S et al. Analyst 2019, 144 (18), 5538–5546; Sloan-Dennison S et al. Chem. Sci. 2019, 10 (6), 1807–1815). In order to further enhance the signal, many studies have focused on changing the shape (Lu X et al. Annual Review of Physical Chemistry 2009, 60 (1), 167–192; Wu X et al. ACS Appl. Mater. Interfaces 2016, 8 (31), 19928–19938; Guo P et al. Nanoscale 2015, 7 (7), 2862– 2868) or size (Guo P et al. Nanoscale 2015, 7 (7), 2862–2868) of the nanoparticles to selectively tune the localized surface plasmon resonance (LSPR) for the highest enhancement. Recently, there has been a growing interest in magnetic nanoparticles as SERS sensors with their ability to magnetically isolate the target of choice from complex media (Donnelly T et al. Chem. Commun. 2014, 50 (85), 12907–12910). Magnetic nanoparticles can help further enhance SERS by inducing aggregation, generating hot spots between particles where coupled plasmon resonance can produce enhancements of around 10 ¹² (Prodan E et al. Science 2003, 302(5644), 419-422; Camden JP et al. Journal of the American Chemical Society. 2008, 130(38), 12616- 12617; Futamata M. Faraday Discuss 2006, 132, 45–61; discussion 85-94). In situ, the magnetic ability to remove them from complex media is helpful. However, there are even more exciting prospects in situ. Magnetic cellular sorting traditionally involves the use of antibodies to specifically target cells or to deplete a heterogeneous cellular population. This concept can be adapted by replacing antibodies with the small peptide mimics targeting specific integrin proteins expressed in the cell to selectively select for high integrin expression. Herein, the ability to differentiate integrins based on their spectral fingerprint is demonstrated. SERS spectra were acquired from Au-coated magnetic nanoparticles functionalized with CDPGYIGSR and cyclic-RGDfC and incubated with purified integrins Į ₄ ȕ ₁ , Į5ȕ1^^Į6ȕ1 DQG^Įvȕ3. These SERS spectra were analyzed using multivariate curve resolution (MCR) to identify the characteristic spectra of each interaction. Furthermore, the confusion matrix for off target interactions was also acquired and compared to the MCR model to ensure signal contributed from off target interactions is insignificant and distinguishable from on target interactions. The ability to distinguish the peptide-protein interactions for 4 integrins that are essential in the process of cellular binding to the extracellular matrix is shown. SW 480 and SW 620 colorectal cancer cell lines were then sorted using both CDPGYIGSR and Cyclic RGDfC functionalized gold coated magnetic nanoparticles to analyze the ability of the particles to select cellular subpopulations. Experimental Chemicals. Absolute Mag™ gold coated magnetic nanoparticles (capped with citrate, 250 nm diameter) were purchased from CD Bioparticles. Cyclic Arg – Gly – Asp-D – Phe – Cys (RGDfC), ^^^^ was purchased from Biosynth. L-Cysteinyl-L-Į-aspartyl-L-prolylglycyl-L- tyrosyl-L-isoleucylglycyl-L-seryl-N5-(diaminomethylene)-L-or nithine (CDPGYIGSR) ^95% was purchased from ThermoFisher Scientific. Recombinant human Integrin Į _v ȕ ₃ ^^^^^^ Recombinant Human ,QWHJULQ^Į ₄ ȕ ₁ 3URWHLQ^^&)^^^^^^^5HFRPELQDQW^+XPDQ^,QWHJULQ^Į ₆ ȕ ₁ 3URWHLQ^^&)^^^^^^DQG^5HFRPELQDQW^+XPDQ^,QWHJULQ^Į6ȕ1^^ &)^^^^^^ZHUH^SXUFKDVHG^IURP^5^'^ systems. Ultrapure water from a Thermo Scientific GenPure UB-TOC/UF xCAD water purification system was used. Magnesium chloride (MgCl ₂ ) and calcium chloride (CaCl ₂ ) were purchased from Sigma Aldrich and phosphate buffered saline (PBS) was bought from Gibco Life Technologies Corporation. RPMI Medium was purchased from ThermoFischer Scientific. Fetal Bovine Serum one shots were purchased from Gibco. SW 480 and SW 620 cell lines were purchased from American Type Cell Culture (ATCC, Manassas, VA, USA). All chemicals were used without further purification. Instrumentation. SERS measurements were performed on an InVia Qontor Renishaw microscope using a 633 nm laser and 50X long working distance objective (0.5 NA). Zetasizer measurements were recorded using a Malvern Zetasizer ZS. Extinction measurements were performed on a Cary 4000 UV Vis instrument. Nanoparticle Tracking Analysis was performed on a Nanosight NS300. An Accument Basic AB 15 pH probe was used for pH measurements. Peptide Functionalization of Gold Coated Magnetic Nanoparticles. To functionalize peptide onto the nanoparticles, 100 μL of gold coated magnetic nanoparticles (1.54 x 10 ¹⁰ nanoparticles/mL) were centrifuged at 7000 relative centrifugal force (rcf) for 15 minutes and the supernatant was removed. To the particles, 630 μL of ultrapure water was added along with 70 μL of stock peptide (1 mM CDPGYIGSR or 1.15 mM Cyclic RGDfC). Particles were then allowed to shake overnight in Eppendorf tubes encapsulated in foil to ensure no photo damage of the nanoparticles occurred. After shaking overnight, the functionalized nanoparticles were centrifuged at 7000 rcf for 15 minutes and the supernatant was removed. The particles were washed with 700 μL of ultrapure water to ensure excess peptide was removed. Particles were resuspended in 1 mL of ultrapure water and the extinction was measured from 200-800 nm. The zHWD^SRWHQWLDO^RI^WKH^SHSWLGH^IXQFWLRQDOL]HG^QDQRSDUWLFOHV^Z DV^WKHQ^UHFRUGHG^DW^S+^§^^^DQG^DIWHU^ DGMXVWPHQW^XVLQJ^^^^^0^+&O^WR^UHDFK^S+^§^^^^3DUWLFOHV^Z HUH^DJDLQ^FHQWULIXJHG^DV^EHIRUH^^DQG^WKH^ supernatant was removed before they were resuspended in 200 μL of ultrapure water and sonicated for about 5 minutes to ensure a stable suspension.5 μL of each peptide nanoparticle suspension was then dropped onto a glass slide and allowed to dry while under an external magnetic force from the bottom of the slide. Raman measurements were then taken using a Renishaw InVia microscope with a 633 nm laser, 0.14 mW of power and 1 second acquisition times. Incubation of Integrin Proteins with Peptide functionalized nanoparticles. Once the gold coated magnetic nanoparticles were functionalized with peptides and characterized, the particles were centrifuged and resuspended in 100 μL of 10% Ca ²⁺ and Mg ²⁺ positive PBS H[FHSW^IRU^&\FOLF^5*'I&^ZLWK^,QWHJULQ^Įvȕ3 which was kept in water. For CDPGYIGSR functionalized nanoparticles, four 50 μL aliquots of nanoparticle VXVSHQVLRQV^ZHUH^VHSDUDWHO\^LQFXEDWHG^ZLWK^^^^^Q0^LQWHJULQ^� �4ȕ1^^^^^^Q0^LQWHJULQ^Į5ȕ1, 373 nM LQWHJULQ^Įvȕ3^^DQG^^^^^Q0^LQWHJULQ^Į6ȕ1, respectively, for 2 hours while shaking. From this solution 5 μL were dropped onto a glass slide while under an external magnetic field and dried in air prior to (SERS) measurements. Protein treatment of cyclic RGDfC functionalized nanoparticles followed a similar procedure. Four different 50 μL aliquots of particles were incubated with 37^^Q0^LQWHJULQ^Į ₄ ȕ ₁ , ^^^^Q0^LQWHJULQ^Į5ȕ1^^^^^Q0^LQWHJULQ^Įvȕ3^^DQG^^^^^Q0^LQ WHJULQ^Į6ȕ1 respectively for 2 hours. The same procedure was used for acquiring SERS spectra. Cell Passage and Sorting. Human SW620 colon cancer cells derived from commercial cell lines (ATCC) were passaged at approximately 80% confluency in 1640 RPMI medium supplemented with 10% Fetal Bovine Serum. The cells were cultured in tissue culture plates at about 80% confluency at 37°C with a 5% CO ₂ humidified atmosphere. Trypsin with EDTA was then added to the flask for 5 minutes to cause detachment from the plate but to maintain viability of the cells prior to recovery and incubation. After passage, cells were allowed to recover their surface markers by sitting for 2 hours with a slight rock in a conical vial every 30 minutes in 10% FBS medium. Afterwards cells were resuspended in 500 μL of DPBS with CaCl2 and MgCl2. Cells were then counted using a 5-fold trypan blue in a Fuchs Rosenthal hemocytometer to determine the number of cells. Enough particles functionalized with peptide also resuspended in DPBS with CaCl2 and MgCl2 were to constitute 40 particles/cell and allowed to incubate for 2 hours with the cells with rocking every 30 minutes. The cells were then counted using a 2-fold dilution with trypan blue to ensure suspension. The cells were then centrifuged for 5 minutes at 300 rcf and the cells were resuspended in 2 mL of EasySep Sorting buffer and allowed to magnetically separate for 5 minutes using an EasySep cell separation magnet. The supernatant was saved as the not magnetically separated (negative) cell concentration. While 2 mL of sorting buffer was used to rinse off the magnetically separated (positive). Each population was once again centrifuged for 5 minutes at 300 rcf and resuspended in 500 μL of DPBS with CaCl2 and MgCl ₂ . Each population was then once again recounted using a 2-fold dilution with Trypan Blue. Data Analysis. MATLAB R2020b was used for all data analysis. Raman spectra from peptides and their interactions with integrin proteins were baseline corrected using adaptive iteratively reweighted penalized least squares (airPLS) (Zhang ZM et al. Analyst 2010, 135 (5), 1138–1146). Spectra were then integrated for total intensity. The standard deviation of the peaks showing consistent low integrated intensity was obtained and a threshold of three times the standard deviation plus the mean of these points were used to set a threshold for spectra that showed signal. The data was reduced to only the spectra that showed any signal above this set threshold (Figure 91-Figure 98). All spectra kept were then normalized to their most intense peak. To reduce the dimensionality of the data, multivariate curve resolution alternating least squares (MCR-ALS) was performed using the PLS Toolbox in MATLAB. The normalized spectra were used without further preprocessing. Results and Discussion Peptide Characterization. Gold coated magnetic nanospheres were functionalized with extracellular matrix (ECM) peptide mimics to investigate particle binding to the extracellular matrix via integrin proteins. For the well-studied fibronectin motif of Arg-Gly-Asp (RGD), cyclic Arg-Gly-Asp-Phe-Cys (RGDfC) has been used previously for similar integrin binding studies (Sloan-Dennison S et al. Chem. Sci. 2019, 10 (6), 1807–1815; Xiao L et al. Anal. Chem. 2016, 88 (12), 6547–6553; Wang H et al. ChemPhysChem 2014, 15 (18), 3944–3949). Figure 68 shows the normalized SERS spectra of cyclic RGDfC functionalized nanoparticles compared to the unfunctionalized citrate capped nanoparticles. The functionalized cyclic RGDfC Nanoparticles exhibit a characteristic peak at 1003 cm ^-1 attributed to phenylalanine, which is not present in stock citrate capped nanoparticles (Cheng WT et al. Microscopy Research and Technique 2005, 68 (2), 75–79; Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195). Large standard deviations as shown in the shaded area (Figure 68) can be explained by the forced aggregation and generation of hot spots from pulling the nanoparticles with a magnet to a spot. While this generates more potential hot spots for increased SERS enhancement, the variable distance between particles can cause the overall intensity to fluctuate drastically (Sergiienko S et al. Physical Chemistry Chemical Physics 2017, 19 (6), 4478–4487). Laminin can be mimicked using the peptide motif of Tyr–Ile–Gly-Ser–Arg (YIGSR). To ensure specific integrin binding and provide thiol chemistry for easy functionalization, the peptide Cys–Asp–Pro –Gly–Tyr-Ile-Gly-Ser-Arg (CDPGYIGSR) was used for its ability to promote cell attachment (Massia SP et al. Journal of Biological Chemistry 1993, 268 (11), 8053– 8059; Maeda T. et al. The Journal of Biochemistry, 1994, 115(2), 182-189). Unlike the RGD peptide motif, there are no readily assigned peaks from peptide on the nanoparticles as shown in Figure 69. However, the change in spectral background from the bare implies a change on the surface of the nanoparticles. Successful functionalization of the peptide to the nanoparticles was confirmed by complimentary methods. Due to the weak acid/base nature of the amino acids, changes in the overall peptide charge can occur as the pH of the solution changes. Thus, the surface charge of the functionalized particles should be responsive to pH, whereas bare particles’ surface charge is less responsive. Zeta potential was measured right after functionalization (pH around 6) and following pH adjustment to 4. These pH’s spans the isoelectric points of the peptides on the particles, 5.4 for Cyclic RGDfC and 5.9 for CDPGYIGSR., respectively. Significant changes in the surface charge are seen when the peptide functionalized particles are protonated at the lower pH compared to the bare citrate capped nanoparticles (Figure 70-Figure 71). Furthermore, for functionalization of particles with both cyclic RGDfC and CDPGYIGSR, a slight plasmon shift from 608 nm to 610 nm is observed and a slight increase in absorbance from the peptide backbone is seen at 205 nm (Figure 88-Figure 89) (Anthis NJ et al. Protein Sci 2013, 22 (6), 851–858). Protein Incubation. It has been previously shown that the interaction between integrin proteins and their corresponding ligands can be studied by analyzing their Raman signals (Sloan- Dennison S et al. Chem. Sci. 2019, 10 (6), 1807–1815; Xiao L et al. Anal. Chem. 2016, 88 (12), 6547–6553; Wang H et al. ChemPhysChem 2014, 15 (18), 3944–3949; Chowdhury MH et al. JBO 2006, 11 (2), 024004). By utilizing the plasmonic features of the gold coated magnetic nanoparticles, a SERS enhancement of the binding site of the ligand-protein interaction can be observed. Extracellular matrix components, and thus peptide mimics, are known to bind to multiple different integrins. This leads to the potential for multiple binding partners on a cell for a given peptide. SERS offers the ability to quickly determine which interaction is occurring, providing information about expression of integrin proteins in a given cell line. This label free approach offers a significant advantage over the commonly performed approaches due to its unique spectral specificity. Figure 72 shows the SERS signals of peptide functionalized nanoparticles with different integrin receptors the attached peptide is known to bind with. The SERS spectra and standard deviations of the known on-target peptide-integrin interactions (Yoshida N et al. Br J Cancer 1999, 80 (12), 1898–1904; Danhier F et al. Mol. Pharmaceutics 2012, 9 (11), 2961–2973; Koivunen E et al. J Cell Biol 1994, 124 (3), 373–380; Sanchez- Esteban J et al. American Journal of Physiology-Lung Cellular and Molecular Physiology 2006, 290 (2), L343–L350; Kapp TG et al. Sci Rep 2017, 7 (1), 39805; Haubner R et al. J. Am. Chem. Soc. 1996, 118 (32), 7461–7472; Liu S. Mol. Pharmaceutics 2006, 3 (5), 472–487; Frith JE et al. Stem Cells Dev 2012, 21 (13), 2442–2456) and reported interactions (CDPGYIGSR – Integrin Į4ȕ1) (Maeda T. et al. The Journal of Biochemistry, 1994, 115(2), 182-189) (Figure 72) offer not only evidence of binding, as a labeled approach does, but also chemical information about the binding site of the interaction. Standard deviations in Figure 72 are also explained by the nature of hot spots (Sergiienko S et al. Physical Chemistry Chemical Physics 2017, 19 (6), 4478–4487), with added uncertainty originating from the size and orientation of these proteins. Even with consistent binding, the portion of the protein in the hot spot might slightly differ, causing some variation in spectral signatures. Prior work shows that a model that characterizes the interaction with a pattern, or spectrum, best representative of that interaction can account for slight variance in the location of the protein relative to the hot spot (Sloan-Dennison S et al. Chem. Sci. 2019, 10 (6), 1807–1815). To identify these spectral patterns, MCR was performed using a 4-component model and was determined sufficient for effectively separating the four unique interactions that were expected and explained 72.6% of the total variance. The loadings for the corresponding components as seen in Figure 73 strongly resemble the average SERS spectra of the average peptide-protein interactions. Additionally, as more components were loaded into the model, the variance explained did not substantially increase after 4 components (Figure 90). Peptide-Protein Discrimination. Scores were generated using each unique peptide integrin binding set. Scores were then plotted in star coordinate plots (Kandogan E et al. Star Coordinates: A Multi-Dimensional Visualization Technique with Uniform Treatment of Dimensions). Star coordinate plots, useful for two-dimensional representation of data greater than two dimensions, were used to visualize the specificity of each component to a specific interaction. These plots display data as a sum of vectors added together. Scores on component 1 were made to be the positive x component, while scores on component 3 were assigned to be negative x. Scores on component 2 were made to be the positive y component for the model, and component 4 were made to be the negative y component. This was to ensure similar peptides would be on opposite sides as to help ensure these components would not create confusion in the 2-dimensional space. The scores on all components were then added together, representing each acquired spectrum, to determine a 2-dimensional coordinate position. Scores for each of the four interaction components from the model in Figure 73 are plotted along a unique vector. The cyclic RGDfC – LQWHJULQ^Į _v ȕ ₃ interaction was placed on the +x axis and to ensure specificity the cyclic RGDfC – LQWHJULQ^Į5ȕ1 was placed on the -x axis of the star coordinate plot. The interactionV^RI^&'3*<,*65^ZLWK^LQWHJULQ^Į ₆ ȕ ₁ ^^\^^DQG^Į ₄ ȕ ₁ (-y) completed the coordinate axes. To determine if the models could effectively be used to distinguish the integrins after they bound to the peptide, spectra from the peptide with and without integrins were scored on the model generated previously. As seen in Figure 74, cyclic RGDfC functionalized particles when ERXQG^ZLWK^LQWHJULQV^Įvȕ3 DQG^Į5ȕ1 QLFHO\^VHSDUDWH^ZLWK^PRVW^RI^WKH^ERXQG^LQWHJULQ^Įvȕ3 scoring RQ^FRPSRQHQW^^^DQG^WKH^ERXQG^LQWHJULQ^Į5ȕ1 scoring along the axis of component 3. The particles in the absence of the integrin cluster near the origin. In Figure 75, a similar behavior is observed for the CDPGYIGSR functionalized particles. While the CDPGYIGSR functionalized particles score near the origin LQ^WKH^DEVHQFH^RI^LQWHJULQ^^WKHLU^LQWHUDFWLRQ^ZLWK^LQWHJULQ^ Į ₆ ȕ ₁ score on FRPSRQHQW^^^DQG^LQWHJULQ^Į ₄ ȕ ₁ score on component 4. This shows that SERS response can differentiate between different proteins binding to the peptides on the nanoparticles. This differentiation indicates SERS spectral data allows for the classification of the integrins without interference of the peptide confounding the signal, and despite the variance in hotspot formation and protein location within these hotspots. To be able to correctly identify the interaction present in cells, as well as to ensure the specificity of the peptide -protein interactions, on-target interactions need to be distinguishable from each other. Using the same model as in Figure 72-Figure 73, a star coordinates plot was made for the 4 on target interactions (Figure 76). Each targeted peptide-protein interaction scatters along a unique separate axis, indicating exclusivity from each other. To ensure that there are no off-target signatures that would affect identification of the peptide-SURWHLQ^LQWHUDFWLRQ^^&'3*<,*65^ZDV^LQFXEDWHG^ ZLWK^LQWHJULQV^Įvȕ3 DQG^Į5ȕ1 and cyclic 5*'I&^ZDV^LQFXEDWHG^ZLWK^LQWHJULQV^Į ₄ ȕ ₁ DQG^Į ₆ ȕ ₁ . As expected, Figure 77 shows that the scores from any off-target interactions that occur scatter randomly towards the center of the plot. This indicates poor scoring on all components of the model, and no consistent interaction. Spectral Analysis. %\^DQDO\]LQJ^WKH^VWUXFWXUH^RI^WKH^5*'^OLJDQG^ZLWK^LQWHJULQV^ Į _v ȕ ₃ (Xiong JP et al. Science 2002, 296 (5565), 151–^^^^^DQG^Į5ȕ1 (Nagae M et al. Journal of Cell Biology 2012, 197 (1), 131–140), spectral assignments were made for the corresponding average SERS spectra. Strong Raman scattering aromatic amino acids such as tyrosine, tryptophan, and phenylalanine largely dominate these signals. )RU^LQWHJULQ^Į ₅ ȕ ₁ peaks at 794 cm ^-1 (Tyr) (Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187– 1195), 1368 (Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195) and 1616 cm ^-1 (Tyr, Trp) (Cheng WT et al. Microscopy Research and Technique 2005, 68 (2), 75–79; Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195; Stone N et al. Journal of Raman Spectroscopy 2002, 33 (7), 564–^^^^^ODUJHO\^GRPLQDWH^WKH^VLJQDO^^)RU^LQWHJULQ^Į _v ȕ ₃ the peaks at 1001 cm ^-1 (Phe) (Cheng WT et al. Microscopy Research and Technique 2005, 68 (2), 75–79; Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195), and 1597 cm ^-1 (Phe) (Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195) dominate. Analysis of the SERS corresponding cyclic RGD-integrin spectra was done by comparing them with their reported crystal structures (Xiong JP et al. Science 2002, 296 (5565), 151–155; Nagae M et al. Journal of Cell Biology 2012, 197 (1), 131–140). With the YIGSR peptide and its corresponding interaction being less studied, it was assumed that amino acids within or near the binding site are represented in the corresponding spectra. Signals from the integrin binding events include peaks that are different than those of the peptides used to target the integrins. Peaks at 611 cm ^-1 (Cys, Arg. C-C Twisting) (Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187– 1195; Chan JW et al. Biophysical Journal 2006, 90 (2), 648–656), 1360 cm ^-1 (Trp) (Cheng WT et al. Microscopy Research and Technique 2005, 68 (2), 75–79; Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195; Stone N et al. Faraday Discuss. 2004, 126 (0), 141–157; Li X et al. Clinical and Biomedical Spectroscopy and Imaging II (2011), paper 808722; Optica Publishing Group, 2011; p 808722), and 1642 cm ^-1 (Amide I) (Stone N et al. Faraday Discuss. 2004, 126 (0), 141–^^^^^GRPLQDWH^IRU^WKH^Į ₄ ȕ ₁ LQWHJULQ^ELQGLQJ^^0HDQZKLOH^LQWHJULQ^Į6ȕ1 shows fewer spectral features although peaks at 1453 cm ^-1 (C-H bending in proteins) (Nogueira GV et al. JBO 2005, 10 (3), 031117) and 1575 cm ^-1 (Trp, His) are present (Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78 (3), 1187–1195; Madzharova F et al. J. Phys. Chem. C 2017, 121 (2), 1235–1242). Full peak assignments for the resulting average spectra of each peptide-protein interaction are included in Table 4-Table 7. Table 4: Peak Identification of Cyclic RGDfC bound to Integrin Į5ȕ1. ¹ Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78(3), 1187– 1195 ² Stone N et al. Faraday Discuss.2004, 126(0), 141–157 ³ Cheng WT et al. Microscopy Research and Technique 2005, 68(2), 75–79 ⁴ Huang Z et al. International Journal of Cancer 2003, 107(6), 1047–1052 ⁵ Bonnier F et al. Analyst 2012, 137(2), 322–332 ⁶ Chan JW et al. Biophysical Journal 2006, 90(2), 648–656 Table 5: Peak Identification of Cyclic RGDfC bound to Integrin Į _v ȕ _3. ¹ Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78(3), 1187– 1195 ² Stone N et al. Faraday Discuss.2004, 126(0), 141–157 ³ Cheng WT et al. Microscopy Research and Technique 2005, 68(2), 75–79 ⁴ Huang Z et al. International Journal of Cancer 2003, 107(6), 1047–1052 ⁷ Lakshmi RJ et al. Radiation Research 2002, 157(2), 175–182 Table 6: Peak Identification of CDPGYIGSR bound to Integrin Į ₄ ȕ _1. ¹ Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78(3), 1187– 1195 ² Stone N et al. Faraday Discuss.2004, 126(0), 141–157 ³ Cheng WT et al. Microscopy Research and Technique 2005, 68(2), 75–79 ⁴ Huang Z et al. International Journal of Cancer 2003, 107(6), 1047–1052 ⁵ Bonnier F et al. Analyst 2012, 137(2), 322–332 ⁶ Chan JW et al. Biophysical Journal 2006, 90(2), 648–656 ⁸ Li X et al. Clinical and Biomedical Spectroscopy and Imaging II (2011), paper 808722; Optica Publishing Group, 2011; p 808722 Table 7: Peak Identification of CDPGYIGSR bound to Integrin Į ₆ ȕ _1. ¹ Zhu G et al. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 2011, 78(3), 1187– 1195 ⁹ Faoláin E et al. Vibrational Spectroscopy 2005, 38, 121–127 ¹⁰ Rix J et al. Journal of The Royal Society Interface 2022, 19 (192), 20220209. Cellular Sorting. SW 480 and SW 620 are commonly used colorectal cancer cell lines that have been well studied in literature and have previously been shown to express integrins binding to Cyclic RGDfC (Xiao L et al. Anal. Chem. 2016, 88 (12), 6547–6553; Wang H et al. ChemPhysChem 2014, 15 (18), 3944–3949). Using the peptide functionalized magnetic nanoparticles, it was found that for Cyclic RGDfC, the percentage of sorted cells were 16.8 ± 6.5% and 11.0 ± 4.0% for SW 480 and SW 620 cell populations respectively as shown in Figure ^^^^3UHYLRXV^OLWHUDWXUH^UHSRUWV^WKDW^6:^^^^^FHOOV^KDYH^EHHQ^ VKRZQ^WR^H[SUHVV^PRUH^RI^WKH^Į _v subunit than the S:^^^^^FHOO^OLQH^ZKLOH^ERWK^KDYH^FRPSDUDEOH^ȕ ₃ expression (Schlaeppi M et al. Cell Adhesion and Communication 1997, 4 (6), 439–455). Also, it is has been previously reported that SW 480 cells have shown higher expression when bound with a CD44+ antibody whiFK^KDV^EHHQ^VKRZQ^WR^KDYH^D^GLUHFW^FRQQHFWLRQ^WKH^,QWHJUL Q^Į ₅ ȕ ₁ (Slater C et al. Oncol Lett 2018, 15 (6), 8516–8526; McFarlane S et al. Oncotarget 2015, 6 (34), 36762–36773). The differences observed using peptide binding interaction based sorting are not statistically significant at the 95% confidence level. Furthermore, when these populations were instead incubated with CDPGYIGSR particles, the percentage of sorted cells 17.1 ± 4.9% and 12.7 ± 4.3% for SW 480 and SW 620 cells respectively as shown in Figure 79. SW 480 cells are UHSRUWHG^WR^KDYH^HTXLYDOHQW^Į ₆ expression (Schlaeppi M et al. Cell Adhesion and Communication 1997, 4 (6), 439–455; Slater C et al. Oncol Lett 2018, 15 (6), 8516–8526; Koretz K et al. Virchows Archiv. 1994, 425, 229-236) but increasHG^ȕ1 expression (Schlaeppi M et al. Cell Adhesion and Communication 1997, 4 (6), 439–455; Koretz K et al. Virchows Archiv. 1994, 425, 229-236). These results indicate the observed differences in the sorting of SW 480 from SW 620 cells are again not statistically significant. Dark field images (Figure 80-Figure 87) were taken of cells fixed with 4% paraformaldehyde to further confirm that the magnetic nanoparticles influenced the sorted cells and ensure binding of the particles was responsible for these observations. More particles are attached to cells in the positively sorted subpopulation than the negatively sorted populations that were not magnetically separated. This could potentially be explained by the fact that most positively sorted cells showed a substantial number of particles compared to the few that would show up in the negatively sorted cells. It should be noted unbound particles were also present in both the positive and negative populations suggesting the number of particles used per cell was excessive and not limiting. Continued investigations may further elucidate differences in receptor expression that regulates the differences in cellular populations sorted. Conclusion. In conclusion, gold coated magnetic nanoparticles were functionalized with extracellular matrix peptide mimics (cyclic RGDfC and CDPGYIGSR), to act as biorecognition agents for specific integrin proteins. SERS, zeta potential, and extinction measurements confirm functionalization of the peptides onto the nanoparticles. The functionalized nanoparticles were then incubated with various integrins that would be present in a cell in situ. Corresponding label free SERS signals were then shown to be specific and were used to differentiate between on- target interactions as well as showing nonspecific binding from off-target interactions. The developed probe and corresponding signals will be useful in potential cellular sorting of heterogeneous cell populations that interact with the extracellular matrix. This also provides a fast, nondestructive method for confirming the presence of integrins in a cellular population. Furthermore, cellular sorting was performed using both peptides on SW 620 and SW 480 cancer cells showing the particles may be useful for potential cellular sorting after further optimization. EXEMPLARY ASPECTS In view of the described compositions, devices, systems, and methods for detection of a target cell using a peptide-modified magnetic particle, herein below are described certain more particularly described aspects of the inventions. The particularly recited aspects should not, however, be interpreted to have any limiting effect on any different claims containing different or more general teachings described herein or that the “particular” aspects are somehow limited in some way other than the inherent meanings of the language and formulas literally used therein. Example 1: An assay comprising: a peptide-modified magnetic particle comprising a particle having a plurality of peptides attached to a surface thereof; wherein each particle comprises a magnetic portion comprising a magnetic material; and wherein each of the plurality of peptides comprises a capture portion configured to capture and bind with at least a first portion of a first target cell; and wherein the capture portion is configured to mimic a first extracellular matrix component. Example 2: The assay of any examples herein, particularly example 1, wherein the particle has an average particle size of from 1 nanometer to 20 micrometers. Example 3: The assay of any examples herein, particularly example 1 or example 2, wherein the particle has an average particle size of from 5 nanometers to 10 micrometers. Example 4: The assay of any examples herein, particularly examples 1-3, wherein the particle has an average particle size of from 5 nanometers to 5 micrometers. Example 5: The assay of any examples herein, particularly example 1 or example 2, wherein the particle has an average particle size of from 5 nanometers to 1 micrometer. Example 6: The assay of any examples herein, particularly examples 1-5, wherein the particle is substantially spherical in shape. Example 7: The assay of any examples herein, particularly examples 1-6, wherein the particle further comprises a plasmonic portion comprising a plasmonic material, such that the particle is a peptide-modified plasmonic magnetic particle. Example 8: The assay of any examples herein, particularly example 7, wherein the plasmonic portion of the particle is configured to enhance a Raman signal of at least a second portion of the first target cell bound to the capture portion of one or more of the plurality of peptides. Example 9: The assay of any examples herein, particularly example 7 or example 8, wherein the particle is a core-shell particle, the magnetic portion comprising the core and the plasmonic portion comprising the shell, such that the plasmonic portion at least partially surrounds the magnetic portion. Example 10: The assay of any examples herein, particularly example 9, wherein the core has an average core size of from 1 nanometer to 20 micrometers. Example 11: The assay of any examples herein, particularly example 9 or example 10, wherein the core has an average core size of from 1 nm to 10 micrometers. Example 12: The assay of any examples herein, particularly examples 9-11, wherein the core has an average core size of from 1 nm to 5 micrometers. Example 13: The assay of any examples herein, particularly examples 9-12, wherein the core has an average core size of from 1 nm to 1 micrometer. Example 14: The assay of any examples herein, particularly examples 9-13, wherein the shell has an average thickness of from 1 nanometer to 5 micrometers. Example 15: The assay of any examples herein, particularly examples 9-14, wherein the shell has an average thickness of from 1 nanometer to 1 micrometer. Example 16: The assay of any examples herein, particularly examples 9-15, wherein the core is substantially spherical in shape. Example 17: The assay of any examples herein, particularly examples 7-16, wherein the plasmonic material comprises a metal selected from the group consisting of Ru, Rh, Pd, Os, Ir, Pt, Au, Ag, Cu, Al, Mg, and combinations thereof. Example 18: The assay of any examples herein, particularly examples 7-17, wherein the plasmonic material comprises a metal selected from the group consisting of Ru, Rh, Pd, Os, Ir, Pt, Au, Ag, Cu, Al, and combinations thereof. Example 19: The assay of any examples herein, particularly examples 7-18, wherein the plasmonic material comprises a metal selected from the group consisting of Pt, Au, Ag, Cu, Al, and combination thereof. Example 20: The assay of any examples herein, particularly examples 7-19, wherein the plasmonic material comprises a metal selected from the group consisting of Au, Ag, and combinations thereof. Example 21: The assay of any examples herein, particularly examples 1-20, wherein the magnetic material comprises a metal selected from the group consisting of Fe, Mn, Ni, Gd, Cu, Co, V, Zn, and combinations thereof. Example 22: The assay of any examples herein, particularly examples 1-21, wherein the magnetic material comprises a metal selected from the group consisting of Fe, Mn, Ni, Gd, and combination thereof. Example 23: The assay of any examples herein, particularly examples 1-22, wherein the magnetic material comprises iron. Example 24: The assay of any examples herein, particularly examples 1-23, wherein the magnetic material comprises an iron oxide. Example 25: The assay of any examples herein, particularly examples 1-24, wherein the magnetic material comprises Fe3O4. Example 26: The assay of any examples herein, particularly examples 1-25, wherein the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle is at least partially dispersed in a solvent. Example 27: The assay of any examples herein, particularly examples 1-26, wherein the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle is disposed on a substrate. Example 28: The assay of any examples herein, particularly examples 1-27, wherein the surface of the particle further comprises a plurality of ligands attached thereto. Example 29: The assay of any examples herein, particularly example 28, wherein the plurality of ligands comprise a second plurality of peptides. Example 30: The assay of any examples herein, particularly example 29, wherein the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle comprises a first population and a second population, the first population being modified with the first plurality of peptides and the second population being modified with the second plurality of peptides. Example 31: The assay of any examples herein, particularly example 29 or example 30, wherein each of the second plurality of peptides has a binding portion configured to: capture and bind to at least a first portion of a second target cell and/or at least a second portion of the first target cell; and mimic a second extracellular matrix component, the second extracellular matrix component being different than the first extracellular matrix component. Example 32: The assay of any examples herein, particularly example 31, wherein the second target cell is different than the first target cell. Example 33: The assay of any examples herein, particularly example 32, wherein the binding portion is configured to selectively bind the first portion of the second target cell. Example 34: The assay of any examples herein, particularly example 31, wherein the binding portion is configured to selectively bind the second portion of the first target cell. Example 35: The assay of any examples herein, particularly examples 1-34, wherein the capture portion of the first plurality of peptides is configured to selectively bind the first portion of the target cell. Example 36: The assay of any examples herein, particularly examples 1-35, wherein the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently has an average length of from 3 to 60 amino acids. Example 37: The assay of any examples herein, particularly examples 1-36, wherein the first extracellular matrix component and the second extracellular matrix component independently comprise a glycosaminoglycan or proteoglycan (e.g., a heparan sulfate, a chondroitin sulfate, a keratan sulfate, perlecan, aggrecan, betaglycan, agrin, neurocan, versican, brevican, decorin, biglycan, testican, bikunin, fibromodulin, lumican, or a combination thereof), a non-proteoglycan polysaccharide (e.g., a hyaluronic acid), a protein (e.g., a collagen, an elastin, a fibronectin, a vitronectin, a laminin, tenascin C, or a combination thereof), or a combination thereof. Example 38: The assay of any examples herein, particularly examples 1-37, wherein the first extracellular matrix component and the second extracellular matrix component independently comprises a hyaluronic acid, a fibronectin, a laminin, a collagen, a glycosaminoglycan, tenascin C, or a combination thereof. Example 39: The assay of any examples herein, particularly examples 1-38, wherein the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises integrin-binding mimetic peptide. Example 40: The assay of any examples herein, particularly examples 1-39, wherein the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises a fibronectin mimetic peptide, a laminin mimetic peptide, a collagen mimetic peptide, a glycosaminoglycan mimetic peptide, a glycosaminoglycan mimetic polysaccharide, or a combination thereof. Example 41: The assay of any examples herein, particularly examples 1-40, wherein the capture portion of each of the first plurality of peptides and/or the binding portion of each of the second plurality of peptides independently comprises a fibronectin mimetic peptide, a laminin mimetic peptide, a collagen mimetic peptide, or a combination thereof. Example 42: The assay of any examples herein, particularly examples 1-41, wherein the capture portion of each of the first plurality of peptides comprises at least 90% identity to RGD, IKVAV, YIGSR, YGYYGDALR, GFOGER, FYFDLR, xNYYSNS, or a combination thereof. Example 43: The assay of any examples herein, particularly examples 1-42, wherein the capture portion of each of the first plurality of peptides comprises at least 90% identity to RGDS, RGDK, RGDF, IKVAV, YIGSR, YGYYGDALR, GROGER, FYFDLR, or a combination thereof. Example 44: The assay of any examples herein, particularly examples 1-43, wherein the capture portion of each of the first plurality of peptides comprises at least 90% identity to cyclic RGDfC, CDPGYIGSR, or a combination thereof. Example 45: The assay of any examples herein, particularly examples 1-44, wherein the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises a cell surface receptor. Example 46: The assay of any examples herein, particularly examples 1-45, wherein the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises a surface protein, an integral protein, a transmembrane protein, or a combination thereof. Example 47: The assay of any examples herein, particularly examples 1-46, wherein the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises an integrin, a cadherin, a tight junction protein, or a combination thereof. Example 48: The assay of any examples herein, particularly examples 1-47, wherein the first portion of the first target cell, the first portion of the second target cell, the second portion of the first target cell, or a combination thereof independently comprises an integrin (e.g., one or more integrins). Example 49: The assay of any examples herein, particularly example 48, wherein the LQWHJULQ^FRPSULVHV^ĮVȕ3^^ĮIIȕ3^^Į5ȕ1^^Į3ȕ1^^Į4ȕ1^^ Į6ȕ1^^Į1ȕ1^^Į2ȕ1, or a combination thereof. Example 50: The assay of any examples herein, particularly examples 1-49, wherein the first target cell and/or the second target cell independently comprises a cancer cell. Example 51: The assay of any examples herein, particularly examples 1-50, wherein the first target cell is a first subpopulation of cancer cells, and the second target cell is a second subpopulation of cancer cells, the first subpopulation and the second subpopulation being different. Example 52: The assay of any examples herein, particularly examples 1-51, wherein the first target cell and/or the second target cell independently comprises a tumor cell. Example 53: The assay of any examples herein, particularly examples 1-52, wherein the first target cell is a first subpopulation of tumor cells, and the second target cell is a second subpopulation of tumor cells, the first subpopulation and the second subpopulation being different. Example 54: The assay of any examples herein, particularly examples 1-53, wherein the first target cell and/or the second target cell independently comprises a cell associated with brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. Example 55: The assay of any examples herein, particularly examples 1-54, wherein the first target cell and/or the second target cell independently comprises a glioblastoma cell or a colorectal cancer cell. Example 56: The assay of any examples herein, particularly examples 1-55, wherein the first target cell and/or the second target cell independently is a glioblastoma cell. Example 57: The assay of any examples herein, particularly examples 1-56, wherein the first target cell is a first subpopulation of glioblastoma cells, and the second target cell is a second subpopulation of glioblastoma cells, the first subpopulation and the second subpopulation being different. Example 58: The assay of any examples herein, particularly examples 1-57, wherein the first target cell and/or the second target cell independently is a colorectal cancer cell. Example 59: The assay of any examples herein, particularly examples 1-58, wherein the first target cell is a first subpopulation of colorectal cancer cells, and the second target cell is a second subpopulation of colorectal cancer cells, the first subpopulation and the second subpopulation being different. Example 60: A method of making the assay of any examples herein, particularly examples 1-59, the method comprising making the peptide-modified magnetic particle or the peptide-modified plasmonic magnetic particle. Example 61: The method of any examples herein, particularly example 60, wherein the method comprises contacting the particle with a plurality of peptides having a functional group configured to covalently or ionically bond to the particle. Example 62: The method of any examples herein, particularly example 61, further comprising making the particle. Example 63: The method of any examples herein, particularly example 61 or example 62, further comprising making the first plurality of peptides having the functional group. Example 64: A method comprising: contacting the assay of any examples herein, particularly examples 7-59 with a liquid sample; subsequently collecting a surface enhanced Raman signal from the liquid sample and the assay; and processing the surface enhanced Raman signal to determine a property of the liquid sample. Example 65: The method of any examples herein, particularly example 64, further comprising applying a magnetic field to the liquid sample and the assay. Example 66: A method comprising: contacting the assay of any examples herein, particularly examples 1-59 with a liquid sample; and subsequently applying a magnetic field to the liquid sample and the assay. Example 67: The method of any examples herein, particularly example 66, further comprising: collecting a surface enhanced Raman signal from the liquid sample and the assay; and processing the surface enhanced Raman signal to determine a property of the liquid sample. Example 68: The method of any examples herein, particularly examples 65-67, wherein applying a magnetic field comprises magnetic activated cellular sorting (MACS). Example 69: The method of any examples herein, particularly examples 65-68, wherein the method comprises repeatedly applying a magnetic field to the liquid sample and the assay. Example 70: The method of any examples herein, particularly examples 65-69, wherein the method comprises serial MACS, e.g. to separate different subpopulations of cells. Example 71: The method of any examples herein, particularly examples 64-70, wherein the property of the liquid sample comprises the presence of the target cell(s) in the liquid sample, the concentration of the target cell(s) in the liquid sample, the identity of the target cell(s), the subtype of the target cell(s), or a combination thereof. Example 72: The method of any examples herein, particularly examples 64-71, further comprising diagnosing and/or monitoring a condition, a disease, or a disorder in a subject based on the property of the liquid sample. Example 73: The method of any examples herein, particularly example 72, wherein the property of the liquid sample is indicative of the condition, the disease, or the disorder. Example 74: The method of any examples herein, particularly examples 72-73, wherein the disease comprises cancer. Example 75: The method of any examples herein, particularly example 74, wherein the cancer comprises brain cancer such as glioblastoma, colorectal cancer, melanoma, lung cancer, gastrointestinal cancers, or a combination thereof. Example 76: The method of any examples herein, particularly example 74 or example 75, wherein the cancer comprises glioblastoma or colorectal cancer. Example 77: The method of any examples herein, particularly examples 74-76, wherein the cancer comprises glioblastoma. Example 78: The method of any examples herein, particularly examples 72-77, further comprising selecting a course of therapy for the subject based on the property of the liquid sample. Example 79: The method of any examples herein, particularly examples 72-78, wherein the method comprises cell subpopulation classification and/or isolation. Other advantages which are obvious and which are inherent to the invention will be evident to one skilled in the art. It will be understood that certain features and sub-combinations are of utility and may be employed without reference to other features and sub-combinations. This is contemplated by and is within the scope of the claims. Since many possible embodiments may be made of the invention without departing from the scope thereof, it is to be understood that all matter herein set forth or shown in the accompanying drawings is to be interpreted as illustrative and not in a limiting sense. The compositions and methods of the appended claims are not limited in scope by the specific compositions and methods described herein, which are intended as illustrations of a few aspects of the claims and any compositions and methods that are functionally equivalent are intended to fall within the scope of the claims. Various modifications of the compositions and methods in addition to those shown and described herein are intended to fall within the scope of the appended claims. Further, while only certain representative method steps disclosed herein are specifically described, other combinations of the method steps also are intended to fall within the scope of the appended claims, even if not specifically recited. Thus, a combination of steps, elements, components, or constituents may be explicitly mentioned herein or less, however, other combinations of steps, elements, components, and constituents are included, even though not explicitly stated.