Do I Need a Standard Curve for Every qPCR Plate?
No, you don't need a standard curve on every plate — but you do need to have generated one at some point for each primer pair, and you need evidence that your assay performance hasn't drifted. The common practice in most relative quantification experiments is to run a standard curve during assay validation, confirm efficiency is between 90–110%, and then use the comparative Ct method (ΔΔCt) for subsequent plates without re-running the curve. This is the Livak method (Livak and Schmittgen, 2001), and it works well when its assumptions are met.
The key assumption: your target and reference gene amplification efficiencies are approximately equal (within ~5% of each other) and both close to 100%. If that holds, you don't need the curve on every plate. If it doesn't — say your GOI amplifies at 87% efficiency and GAPDH at 99% — you either need to re-optimize your primers or switch to the Pfaffl correction (Pfaffl, 2001), which uses the actual efficiency values from your standard curves. In that case, you still don't need the curve on every plate, but you need a reliable efficiency estimate, and you should re-verify it periodically.
When You Can Skip the Standard Curve
For routine relative quantification with well-validated assays, running a standard curve on every plate is a waste of wells and cDNA. Here's what "well-validated" actually means:
- You've run a 5-point, 4-fold (or 5-fold) dilution series at least once — ideally twice, on different days — for each primer pair.
- Efficiency is 90–110% (slope between –3.58 and –3.10) with an R² ≥ 0.98.
- Your target and reference efficiencies are within ~5 percentage points of each other if you're using ΔΔCt.
- Your NTCs are clean — no amplification, or Ct > 38 with a melt curve that doesn't match your product.
- You're using the same reagent lot, primer stocks, and instrument. When any of these change, re-validate.
If all five boxes are checked, you can run plates with just your experimental samples, calibrator/control samples, reference genes, and NTCs. Most published qPCR studies work this way, and the MIQE guidelines (Bustin et al., 2009) support it — they require that efficiency be reported, not that a curve appear on every plate.
A practical middle ground that many labs use: include an inter-run calibrator (IRC) on every plate. This is a single concentration of a sample that you aliquot in bulk, freeze, and run in duplicate or triplicate on every plate. It doesn't give you a full efficiency estimate, but it catches catastrophic problems — a bad reagent mix, a miscalibrated thermal block, degraded primers. If your IRC Ct drifts by more than 0.5–1.0 Ct from its historical mean, something has changed and you should investigate before trusting the rest of the plate.
When You Absolutely Need a Standard Curve
There are situations where the curve isn't optional:
Absolute quantification. If you need to report copy numbers — viral load, transgene copies per genome, bacterial burden — you need a standard curve of known concentration on every plate (or at least within the same run batch with an IRC linking plates together). Your result is only as good as the curve it's interpolated from. Use plasmid DNA, gBlocks, or linearized vector with a well-quantified stock. Verify the concentration by at least two methods (spectrophotometry + fluorometry with Qubit or PicoGreen).
New primer pairs. Until you've characterized the assay, you have no basis for assuming anything about efficiency. Run the full curve. If your R² is below 0.98 or efficiency is outside 90–110%, troubleshoot before generating any experimental data. Common fixes: adjust primer concentration (try 200, 300, and 400 nM), optimize annealing temperature (gradient from 56–64°C on a CFX96 or QuantStudio), or redesign primers that land in a region with less secondary structure.
Multiplex TaqMan assays. When you're running two or more probe-based assays in the same well, competition for reagents can shift efficiency in ways that aren't predictable from singleplex curves. Validate the multiplex efficiency for each target at the concentrations you'll actually use. If you see efficiency drop below 90% in multiplex, reduce the primer/probe concentration of the high-abundance target (often the reference gene) to 100–150 nM to give the rarer target more room.
When reagents, instruments, or primer lots change. A new lot of PowerUp SYBR or Luna Universal qPCR Mix might perform identically — or it might not. Same for a fresh primer resuspension. Re-run the curve. It takes one plate and saves you from weeks of confusing data.
How to Generate a Standard Curve You Can Actually Trust
A surprising number of efficiency problems trace back to how the curve was made, not to the assay itself.
Use a relevant template. For cDNA-based assays, make your standard curve from pooled cDNA, not from plasmid or gDNA. Pooled cDNA from your experimental samples captures the actual matrix your primers will encounter — reverse transcription artifacts, background complexity, and all. For gDNA-based assays (copy number, ChIP-qPCR), use gDNA from a relevant cell type.
Use sufficient dilution points. Five points is the minimum for a reliable linear regression. Four-fold or five-fold serial dilutions work well because they space your Ct values about 2 cycles apart, giving good resolution across the range. A typical series: undiluted pooled cDNA, then 1:4, 1:16, 1:64, 1:256. If your undiluted sample gives a Ct around 18 for a moderately expressed gene, the 1:256 dilution will be around Ct 26 — well within the reliable detection range.
Pipet carefully. This sounds patronizing, but serial dilution errors are the single most common reason for bad standard curves. Use fresh tips at every step. Vortex or flick-mix each dilution before taking the next aliquot. A 5% pipetting error at each step compounds across the series and will distort your slope. If your standard curves routinely show R² values of 0.96–0.97, the problem is almost certainly technique, not biology.
Run technical replicates. Duplicates are the minimum; triplicates are better for the standard curve specifically. If one replicate in a triplicate set deviates by more than 0.5 Ct from the other two, exclude it — it's a pipetting outlier.
The Efficiency Math, Briefly
The slope of your standard curve (log₁₀ of dilution factor vs. Ct) relates to efficiency by:
E = 10^(–1/slope) – 1
A perfect doubling every cycle gives a slope of –3.322 and E = 1.00 (100%). In practice:
| Slope | Efficiency |
|---|---|
| –3.10 | 110% |
| –3.32 | 100% |
| –3.58 | 90% |
| –3.92 | 80% |
Anything outside 90–110% means something is off — primer dimers competing at low template concentrations, inhibitors in your cDNA, secondary structure in the amplicon, or suboptimal primer design. An efficiency above 110% usually indicates inhibition at high concentrations (the concentrated points are artificially delayed, steepening the apparent slope) or pipetting error.
If your target and reference efficiencies differ by more than 5 percentage points but are both within an acceptable range, use the Pfaffl model instead of ΔΔCt. The Pfaffl equation accounts for unequal efficiencies:
Ratio = (E_target)^ΔCt_target / (E_ref)^ΔCt_ref
where ΔCt is (control – treated) for each gene. This gives you accurate fold changes without forcing the equal-efficiency assumption.
What About Plate-to-Plate Variability?
This is the real concern behind the "standard curve on every plate" question. Plate-to-plate variability is real, especially on instruments where individual wells can vary by 0.3–0.5 Ct due to edge effects or optical inconsistencies (I'm looking at you, older 96-well block instruments). The QuantStudio 5/7 and CFX Opus have improved on this, but it's still not zero.
The best practices to manage this without burning 15 wells on a standard curve every time:
- Use the same well positions for your calibrator/control samples across plates.
- Include an inter-run calibrator (2–3 wells) on every plate.
- Don't split biological comparisons across plates if you can avoid it. If you're comparing treated vs. untreated, try to get both conditions on the same plate. If you must split across plates, the IRC lets you apply a plate correction factor.
- Avoid edge wells (A1, A12, H1, H12 on a 96-well plate) for critical samples. Use them for NTCs.
The Short Version
Validate each primer pair with a proper standard curve once. Confirm efficiency is 90–110% and efficiencies are matched if using ΔΔCt. After that, run your experimental plates with an inter-run calibrator and NTCs — no standard curve needed. Re-validate when reagents, instruments, or primer lots change, or if your IRC Ct starts drifting.
If you're running a lot of plates and want to track efficiency validation, IRC consistency, and replicate agreement without building your own spreadsheet system, VoilaPCR handles all of this automatically — flag it, plot it, and keep a record so you know exactly when your last validation was and whether your current data are trustworthy.