ΔΔCT Fold Change Calculator

The ΔΔCT Method The ΔΔCT (delta delta CT) method calculates relative gene expression between samples. Developed by Livak and Schmittgen (USA, 2001), it is the most cited qPCR analysis method there is. Formula: Fold change = 2^(−ΔΔCT)

Step-by-Step Calculation

1. ΔCT (sample) = CT(target gene) − CT(reference gene) [normalizes for RNA input]
2. ΔCT (control) = CT(target gene, control) − CT(reference gene, control)
3. ΔΔCT = ΔCT(sample) − ΔCT(control)
4. Fold change = 2^(−ΔΔCT)

Interpretation Fold change > 1: gene is upregulated in treatment vs control. Fold change < 1: gene is downregulated. Fold change = 1: no change in expression. Fold change = 2: gene is expressed twice as much in the treated sample. Fold change = 0.5: gene is expressed at half the level.

CT Values CT (cycle threshold) = number of PCR cycles needed to cross the fluorescence threshold. Lower CT = more template (more gene expression). Each 1 CT difference = ~2-fold difference in abundance (assuming 100% efficiency). Typical CT range: 15 to 35. Above 35 the reading is unreliable for most assays, because you are down to a handful of template molecules and Poisson noise starts to dominate.

Reference Genes (Housekeeping Genes) Must be stably expressed across all experimental conditions. Common choices: GAPDH, β-actin, HPRT1, 18S rRNA, RPLP0. Validate reference gene stability before using (with NormFinder or geNorm tools). Using an unstable reference gene invalidates the entire dataset.

Assumptions This method assumes ~100% PCR efficiency (E ≈ 2). With efficiency E: Fold change = E^(−ΔΔCT). Use corrected formula if efficiency is measured via standard curve. Biological replicates (n ≥ 3) are required for statistical significance.

Worked example

The placeholder values on this page are a complete run. Treatment sample: target CT 22.5, reference CT 18.3. Control sample: target CT 25.1, reference CT 18.2.

• ΔCT (treatment) = 22.5 − 18.3 = 4.2
• ΔCT (control) = 25.1 − 18.2 = 6.9
• ΔΔCT = 4.2 − 6.9 = −2.7
• Fold change = 2^2.7 = 6.4980, so log₂FC = 2.7

The gene is roughly 6.5 times more abundant in the treated sample. The reference gene came in at 18.3 against 18.2, a difference of 0.1 cycles, which is the reassuring part: it means the two RNA inputs were comparable and the 2.6-cycle shift in the target is real rather than a loading artifact.

Now run the same numbers at 90% efficiency instead of 100%. E becomes 1.9 rather than 2.0, and the fold change falls to 1.9^2.7 = 5.6576. A 10-point efficiency error moved the answer by about 13%, which is why the standard curve is worth running once per primer pair.

What counts as a real change

A fold change of 1.3 is not a finding. Technical variation alone routinely produces 1.2 to 1.5 between replicate wells, and biological variation between animals or plates is larger still.

The usual working threshold is 2-fold in either direction, meaning log₂FC of ±1, backed by at least three biological replicates and a statistical test on the ΔCT values rather than on the fold changes themselves. Fold change is a ratio, so it is not normally distributed and a t-test on it is not valid. Run the test on ΔCT, then report the fold change.