Traditionally, scientists have used the technique of the Polymerase Chain Reaction (PCR) to synthesise defined sequences of DNA. This generally involves a three step procedure: separation of the DNA to be amplified (template DNA); annealing of short complimentary DNA sequences (primers) to the template DNA and finally the addition of deoxynucleotides to the primer strands in order to copy the template DNA. This is usually performed in a thermal cycling machine where a cycle of three different temperatures is repeated approximately 25-35 times. Template DNA separation and synthesis steps occur at defined temperatures. However, the temperature at which the primer binds to the DNA, may need optimising in order for this step to occur efficiently and achieve desirable PCR results. Primer annealing optimisation experiments usually involve setting up a number of different experiments where only the primer annealing temperature is varied. The experiment may need to be performed 3 or 4 times in order to determine the optimum binding temperature. These experiments would have to be repeated each time a new set of primers was required for different PCRs. The development of a temperature gradient block enables the scientists to determine the optimum binding temperatures of several primer sets in a single experiment.

However, prior art derives which produce a thermal gradient have been relatively complex and expensive.

The present invention seeks to provide an improved thermal cycler.

According to an aspect of the present invention, there is provided a thermal cycler including a sample plate able to

hold a plurality of samples, a plurality of heating and cooling elements located along the sampling plate, supply means operable to provide a current to the heating and cooling elements, and switching means able to switch current individually through each heating or cooling element or each of a plurality of sets of heating or cooling elements.

The temperature gradient block can achieve a set of discrete temperatures simultaneously, thereby decreasing the number of optimisation experiments the scientist needs to perform and therefore the workload.

An embodiment of the present invention is described below, by way of example only, with reference to the accompanying drawings, in which: Figure 1 is a perspective view of an embodiment of heating and cooling block of a thermal cycler; Figure 2 shows different elevational views of the block of Figure 1 with the sample plate attached and removed to reveal a plurality of heating and cooling elements; Figures 3 to 8 show an example of circuit diagram for driving the heating and cooling elements of Figure 2 in different manners; and Figure 9 is a graph showing the time/temperature characteristics of the thermal cycler of Figure 1.

Controlled heating and cooling of biological samples is preferably achieved by using Thermoelectric Coolers (TECs).

This embodiment uses 8 such devices, positioned between a metal sample plate and a heat sink 2 arranged a shown in Figure 1.

A TEC will generate a temperature differential across its surfaces when a current is applied. By altering the polarity of the applied current, the temperature of the sample plate can be raised or lowered, as is known in the art. When uniform heating and cooling of the sample plate is required, all 8 TECs are connected in series and in the same polarity.

To produce temperature gradients across the sample plate, the relative polarity of individual TECs is changed or they are bypassed altogether.

Referring to Figure 2, a circuit board 3 (described in greater detail below) is located on the heat sink, and in this example, eight TECs 4 are located over the circuit board 3. The sample plate 1 is of conventional form and typically is able to hold many samples in an organised array.

Referring to Figures 3 to 8, the circuit used to deliver current to the TECs 4 is shown in various operating states.

The specific component types mentioned below are given by way of example only. The skilled person will readily be able to find alternatives.

Current flow through the TECs is switched by means of MOSFETs (International Rectifier IRLI3705N) shown in Figure 3 as Q1-Q6. The design uses International Rectifier PVI5100 Photovoltaic Isolators, U1-U6, to provide a gate drive voltage for each MOSFET which is isolated from the control interface.

To achieve uniform heating and cooling, control signal C2 is set to logic"1"to provide base current for Q12 which activates U3 and U5 and enables MOSFETs Ql and Q4. With a positive voltage on A with respect to B, current flow in the TECs is as shown in Figure 4, resulting in uniform cooling of the sample plate 1. With a negative voltage on A with respect to B, current flow in the TECs is as shown in Figure 5, resulting in uniform heating of the sample plate. In this

case, current flows through the parasitic diodes in Ql and Q4.

To achieve a temperature gradient across the sample plate 1, control signal Cl is set to logic l'to provide base current for Q7 which activates U2 and Ul and enables MOSFETs Q2 and Q6. With a positive voltage on A with respect to B, current flow in each of the TECs is as shown in Figure 6, resulting in a cooling effect in TECs Pl, P2, P3 and P4 and a heating effect in TECs P5, P6, P7 and P8. Note that in this case, current flows through the parasitic diode in Q3.

In order to obtain a more linear gradient across the sample plate 1, less heating/cooling is required from the inner TECs P3, P4, P5 and P6. This is achieved by bypassing these TECs whilst keeping TECs Pl, P2, P7, and P8 in a gradient' configuration. Control signals Cl and CO are set to logic '1'tu provide base current for Q7 and Q13 which activates Ul, U2, U4 and U6, and enables MOSFETs Q2, Q3, Q5 and Q6.

With a positive voltage on A with respect to B, current flow in the TECs is as shown in Figure 7, resulting in a cooling effect in TECs P1 and P2 and a heating effect in TECs P7 and P8. Note that in this case current flows through the parasitic diode in Q3.

By setting control signals C2 and CO to logic l'base current is provided for Q12 and Q13 which activates U3, U4, U5 and U6 and enables MOSFETs Ql, Q3, Q4 and Q5. With a negative voltage on A with respect to B, current flow in the TECs as shown in Figure 8, resulting in a heating effect in TECs Pl, P2, P3 and P4 only. This mode of operation is used to help re-establish temperature uniformity across the sample plate. Note that in this case current flows through the parasitic diode in Q2.

It will be apparent that the current through the inner TECs P3, P4, P7 and P8 can be switched intermittently to produce

any desired temperature gradient, as can be done with the outer TECs Pl, P2, P5 and P6. The duty cycles suitable for the various TECs can be determined by experimentation.

Figure 9 shows a typical temperature/time characteristic obtained using this technique. Each line on the graph represents the temperature of particular locations on a line running from left to right across the sample plate 1.

Zone A represents uniform cooling or heating. Current flow through the TECs is as shown in Figure 4 (cooling) and Figure 5 (heating).

Zone B represents operation in gradient mode. In this mode, uniform heating and cooling is used in combination with the gradient operation, represented in Figure 6 and 7, to achieve the desired temperature set-point and the desired temperature gradient.

Zone C represents the recovery from gradient mode.

Temperature uniformity across the sample plate is re-established by a combination of uniform heating and partial heating represented in Figure 8.

It will be apparent that the pairing of TECs for the purposes of heating/cooling is merely exemplary.

A description of the preferred method of operation of the system, advantageously provided in software, is given below.

The eight Peltier Thermo Electric Coolers (TECs) are wired in four pairs of two. Three control signals are used to switch the arrangement of Field Effect Transistors (FETs) Ql to Q4, which in turn are used to switch pairs of TECs in or out of the circuit. The four pairs of TECs can be driven in the following modes:

1. Common mode, where all four pairs are used for heating or cooling.

2. Differential mode, where two pairs at one end are used for heating and two pairs at the opposite end are used for cooling. This mode is used during a gradient step.

3. Half Differential mode, where the two outer pairs only are driven. This mode is used to improve linearity across the block during a gradient step.

4. Start Gradient mode, where the two pairs used at the cooler end of the block are cooled, and the other two pairs are not driven. This mode is used when starting a gradient step.

5. Stop Gradient mode, where the two pairs at the cooler end of the block are heated, and the other two pairs are not driven. This mode is used when ending a gradient step.

A PID loop controls the temperature of the block, based on measurements taken from a thermistor embedded in one end of the block. Switching the TECs between the various operating modes for a calculated period of time enables the temperature gradient to be accurately controlled.

FET States The Gradient Temperature Block circuitry gives nine possible FET states. These states are the combination on the three control bits C2, Cl and C0, with the power supply polarity, + or-.

Each of the sixteen combinations are named 0+, 0-, 1+, 1- .... 7+, 7-.

Of these sixteen combinations, only nine are usable because the FETs have diodes allowing reverse current to flow: NAME 1 2 3 4 STATES Off----0-, 1- Heat H H H H 0+, 2+, 4+, 6+ Cool C C C C 4- Differential C C H H 2- StartGradient C C--6- Stop Gradient H H--1+, 3+, 5+, 7+ Half Differential C--H 3- CoolEnds C--C 5- Not Used ( !) C c h c 7-

Column 1 represents the variable (cooler) end of the block, column 4 is the fixed end.

H indicates Heating, C is Cooling and-is no change Mode 7-is not used because it gives c=1/3 Cooling and h=1/3 Heating Only the states in bold are used in the Express program State 6-is only on for one time slot, and has a negligible effect. It is thought that this method of starting up a gradient was'never fully implemented.

A 50: 50 ratio mix of 2-and 3-gives an average as follows: 2-C C H H 3-C--H Sum 2C C H 2H This gives a linear slope State 3+ indicates the C2=0, Cl=CO=l and the power supply is in Positive mode.

This corresponds to the +ve supply connected to the B terminal on the circuit diagram.

Mode of Operation A global variable called gradient is set when the temperature gradient needs to be changed. This is calculated from the measured gradient and required gradient.

In Normal mode CurrentDemand (O) is used to set the power supply.

In Gradient mode CurrentDemand (1) is used to set the power supply. This is clipped to 0 or negative values since the Differential, Half Gradient and Reverse Differential FET states only work with one direction of current.

In Normal modes, one time value, called ZoneTime (O), is calculated from the CurrentDemand (O) value.

In Gradient modes, two time values are calculated from the CurrentDemand (1) value. These are called ZoneTime (O) and ZoneTime (1).

ZoneTime (1) is always set to one half of ZoneTime (O).

ZoneTime (O) is the number of time slots out of 190 which the power supply is on.

ZoneTime (1) marks the mid point of ZoneTime (O) so that in increasing gradient mode the FET state can be 50% differential and 50% half differential (ends only).

A finite sate machine called UpdateCurrentDemand is used to control the process- State 0 initialises variables State 1 decides on one of three sub-machines: 1. Normal

2. Increase the gradient 3. Decrease the gradient For an increase or decrease in gradient 256/446 slots are Normal mode and 190 are Gradient mode.

Normal Mode: The process is divided up into 190 timeslots.

The FET state is switched to Normal.

The power supply is on for ZoneTime out 190 slots, then switched off.

Increasing Gradient Mode: The process is divided up into 190+256=446 timeslots.

The FET state is Differential (2-) for half ZoneTime (O), then Half Differential (3-) for another half ZoneTime (O). (This gives a more linear slope than just one of the states on its own).

Then the power supply is switched off until slot 191.

Then the FET state is switched to Normal (4+, 4-) and Gradient mode disabled for 258 slots.

Decreasing Gradient Mode: The process is divided up into 190=266=446 timeslots.

The FET state is Reverse Differential (5+) for ZoneTime (O).

The FET state is then switched to Normal and Gradient mode disabled for 256 slots.

Then the power supply is switched off until slot 191.

Then the FET state is switched to Normal (4+, 4-) and Gradient mode disabled for 256 slots.

For either gradient mode the power supply is on for ZoneTime in gradient mode (if required) out of the first 190 slots, then for another ZoneTime out of the next 256 slots in normal mode.

Summary In normal (non gradient) mode the power supply is pulse width modulated so that it is on for ZoneTime slots out of 190.

The PWM ratio is calculated by the main PID loop.

In gradient mode, when a change of gradient is required, the FET state is adjusted to give the required change, and the gradient PID loop calculates the ZoneTime (O), ZoneTime (1) and required current.

When the 190 slots are completed, the unit is switched into normal mode for 256 slots, so that the overall block temperature can be adjusted. If a change in gradient is not required the unit remains in normal mode.

The disclosures in British patent application no. 9826237.1, from which this application claims priority, and in the abstract accompanying this application are incorporated herein by reference.