Runs in your browser Report an error © 2026 Medecsta
RT-qPCR Analysis All research tools

RT-qPCR Analysis

Plate data

Paste from Excel or from the instrument's export, header row included. Needed columns: Sample, Target, Cq (or Ct). Optional: Biological Set Name (or Group), Content (NTC, NRT), Well. Other columns are ignored. A well with no Cq is left out, never counted as a number.

Reference gene(s)
Paste a table to list its genes.

No analysis yet

Paste a Cq table, tick the reference gene, choose the control group and press Analyse. The simulated example shows a full analysis with a known answer.

Fold change

By group

Fold change is the geometric mean, 2^(-mean ΔΔCq); its range is 2^(-(mean ± SD)). Mean, SD and SEM are of the samples' individual fold changes. n is the number of samples, not wells.

By sample

Quality checks

    Controls

    Reference gene stability

    Technical replicates

    The tool flags a well and never removes one. To leave a well out, delete its row in the Cq table and analyse again.

    Primer efficiency from a dilution series

    Needed columns: Target, Quantity (or Starting Quantity), Cq. Give relative amounts that fall as the sample is diluted (1, 0.2, 0.04), not dilution factors. Where a Content column marks standards (Std), only those wells are fitted.

    A straight line through Cq against log10(quantity), fitted to every well. E = 10^(-1/slope); efficiency % = (E - 1) × 100. Accepted: 90 to 110 %, with R² of 0.98 or higher.

    Methods and references

    • Technical replicates. The wells of one sample for one gene are averaged; their SD (n - 1) and range are reported. A well with no Cq (NaN, N/A, Undetermined, blank) is left out and counted. It is never replaced by 0, 40 or any other number.
    • Amplification factor. E = 1 + efficiency % / 100, so E = 2 at 100 %.
    • ΔCq. Cq(target) × log2 E(target) minus the mean, over the reference genes, of Cq(reference) × log2 E(reference). With every efficiency at 100 % this is the ordinary ΔCq = Cq(target) - Cq(reference). With measured efficiencies it is on the log2-quantity scale.
    • ΔΔCq and fold change. ΔΔCq = ΔCq(sample) - mean ΔCq of the control group's samples. Fold change = 2^(-ΔΔCq), so the control group's geometric mean is 1.
    • One formula, three methods. At 100 % with one reference gene it is the 2^(-ΔΔCq) method (Livak and Schmittgen 2001). With measured efficiencies it is Pfaffl's ratio (2001). With several reference genes the target is divided by the geometric mean of their relative quantities (Vandesompele 2002; Hellemans 2007). CFX Maestro's normalized expression with a control selected follows the same model (User Guide 10000126764, Appendix A).
    • Groups. Expression is log-normal, so a group is summarised on the ΔΔCq scale: geometric mean fold change 2^(-mean ΔΔCq), range 2^(-(mean ± SD)). The arithmetic mean, SD and SEM of the individual fold changes are given as well, because many papers plot those. A sample without a group is its own group.
    • Per-sample range. SD of ΔCq = square root of the sum of the squared replicate SDs (target, and reference divided by the number of reference genes); range = 2^(-(ΔΔCq ± SD)). The calibrator is treated as a constant (Applied Biosystems User Bulletin #2).
    • Efficiency. Linear regression of Cq on log10(quantity) over every well. E = 10^(-1/slope); efficiency % = (E - 1) × 100.
    • Limits. Replicate wells within 0.5 Cq. No-template control: no Cq, or later than 35. No-RT control: no Cq, or at least 5 cycles later than its sample. Reference gene: SD of Cq across all samples no more than 0.5. Efficiency 90 to 110 % and R² at least 0.98. These are the acceptance limits of the SYBR Green lab protocol this tool was built from; they are conventions, not laws, and a journal or kit may ask for others. Two limits are this tool's own: a standard curve needs at least 3 dilution levels, and a sample with a mean Cq later than 35 is flagged.
    • Not done. Significance tests, automatic outlier removal, inter-run calibration between plates, absolute copy number. Paste one run at a time unless the plates need no calibration.
    • Checked against. The worked table in User Bulletin #2 (c-myc and GAPDH in four tissues), a hand-worked Pfaffl ratio, and a second calculation written in a different order and compared on random plates: qpcr.test.mjs. Not yet compared with CFX Maestro output on a real plate.
    • Example data. Simulated, not from any study: example-plate.csv and example-standard-curve.csv, written by make-example.mjs. Built in: efficiencies Gapdh 98 %, Cyp7a1 95 %, Tnf 103 %; fold change against Control of 0.25 and 0.5 (Cyp7a1) and 4 and 2 (Tnf) in the Model and Treated groups. One Tnf well of Model 2 is 0.85 Cq late on purpose: it is flagged, and it pulls the Model fold change of Tnf below 4 until its row is deleted. One Tnf no-template well has a Cq of 37.42.

    Livak KJ, Schmittgen TD. Methods. 2001;25(4):402-408. PMID 11846609.
    Pfaffl MW. Nucleic Acids Res. 2001;29(9):e45. PMID 11328886.
    Vandesompele J, et al. Genome Biol. 2002;3(7):RESEARCH0034. PMID 12184808.
    Hellemans J, et al. Genome Biol. 2007;8(2):R19. PMID 17291332.
    Bustin SA, et al. The MIQE guidelines. Clin Chem. 2009;55(4):611-622. PMID 19246619.
    Applied Biosystems. User Bulletin #2: ABI PRISM 7700 Sequence Detection System. 1997, updated 2001.
    Bio-Rad Laboratories. CFX Maestro Software User Guide, 10000126764.

    Runs in your browser — nothing is uploaded In review — research reference, not a diagnostic test. Check anything that matters and report anything that looks wrong.