PCR setup is the kind of task automation was made for and the kind of task that punishes automation done carelessly. It is repetitive, it is high volume, and by hand it is tedious enough that a technician setting up a full plate late in the day will eventually miss a well or double one. A liquid handler never gets bored. But PCR is also unforgiving in ways that hide until the amplification curve comes back wrong, and the two things that most often go wrong are precisely the two things that look trivial: the volumes are tiny, and a single stray molecule can ruin everything.
So automating PCR is not really about moving liquid faster. It is about protecting two properties at once: the ratio between master mix and template, which small-volume error corrupts, and the purity of every well, which amplicon carryover destroys. Everything below is in service of one or the other.
The ratio-fidelity problem
A PCR reaction is a recipe, and the recipe is a ratio. When you are dispensing two to five microliters of template into a master mix, an absolute error of a few hundred nanoliters is a large fraction of the intended volume, and it shifts the ratio enough to matter. In endpoint PCR that can mean a weak or failed band. In qPCR it moves the Cq value, and because Cq is read on a log scale, a volume error you would shrug off elsewhere becomes a quantification error you cannot. Precision at these volumes is not a nicety, it is the entire assay.
That means the low-volume behaviour of the class is what you tune first: the settling delay that lets a tiny slug of liquid actually leave the tip, the aspiration and dispense speeds slow enough that the volume is real and not partly air, and a dispense position that puts the liquid where it will mix rather than clinging to the tip. At two microliters the difference between a class that was tuned and one that was assumed is the difference between a clean standard curve and a noisy one.
Master mix is not water
Distributing master mix is the bulk of the work, and master mix behaves nothing like the buffers people quietly benchmark their classes against. Most contain glycerol, from the polymerase storage buffer, and glycerol makes the mix viscous. A viscous liquid resists being drawn up, so a speed tuned for water will under-aspirate, and it resists leaving the tip, so it needs time to catch up.
- Slow aspiration: draw the mix up slowly enough that the liquid keeps pace with the plunger, because a viscous mix pulled too fast arrives short and full of air.
- Aspiration and dispense delays: pause with the tip in place after each move so the viscous column finishes flowing before the tip lifts, rather than trailing liquid up the outside.
- Pre-wetting: cycle the mix in and out of the tip a few times before the real aspiration so the interior wets out and the first delivered volume matches every one after it.
Get these right and one aspiration can feed a row of reactions with volumes that match end to end. Get them wrong and the first well and the last well of the plate carry different amounts of enzyme, which is a gradient you will spend an afternoon failing to explain.
Contamination control is the whole game
The thing that makes PCR contamination uniquely nasty is that the reaction amplifies its own mistakes. A few molecules of yesterday amplicon carried into today plate get copied billions of times right alongside your real target, and the result is a false positive that looks exactly like a real one. There is no dispense parameter that fixes this after the fact, so contamination control is designed into the workflow rather than tuned into a class.
- Filter tips: use barrier tips throughout so aerosols cannot reach the channel and travel between samples inside the instrument.
- Fresh tips per sample: never reuse a tip across templates, because the tip is the single most direct path for one sample to contaminate the next.
- Physical separation: keep pre-PCR setup and post-PCR handling in different areas, ideally different rooms, so amplified product never shares air or surfaces with the reactions being built.
- Unidirectional workflow: move reagents and people from clean to dirty and never back, so nothing that has seen amplified product returns to the setup deck.
- Enzymatic guards where used: some workflows add uracil handling so that any carried-over product from a previous reaction is degraded before amplification begins.
None of these are settings you dial in and forget. They are constraints on how the deck is laid out and how the run is ordered, and the liquid handler helps mostly by being consistent: fresh tips every time, the same clean path every run, no tired shortcuts at the end of a plate.
Dispensing, mixing, and the qPCR extras
Where you place the liquid matters as much as how much you place. Dispensing template to the bottom of the well or against the wall, then mixing it into the master mix with a few gentle aspirate-dispense cycles, gives you a homogeneous reaction rather than a bead of template sitting on top of the mix. The mixing has to be gentle, because the same energy that homogenises the well also entrains air.
That is where qPCR raises the bar. The reaction is read optically, and a bubble sitting in the light path scatters signal and corrupts the read, so a dispense that would be fine for endpoint PCR can quietly fail qPCR by leaving air behind. The class for a qPCR plate has to dispense without entraining air and mix without whipping bubbles in, which usually means slower still and a dispense that breaks the surface cleanly rather than plunging.
Two more details close the loop. Enzymes are temperature sensitive, so hold reagents on a cold block through setup rather than letting a viscous mix warm and change behaviour mid-run. And because the volumes are small, plan for dead volume and for evaporation across a plate that may sit open while every well is filled; seal promptly and consider the edge wells, which dry fastest. The failure modes here are a short list and every one of them is avoidable: skewed ratios from sloppy small-volume dispensing, contamination from a broken clean path, and bubbles from a dispense tuned for volume rather than for an optical read.
A liquid handler will set up a PCR plate perfectly a thousand times or ruin it a thousand times. The difference is entirely in whether the class respects that the volumes are tiny and the contamination is permanent.