threshold cycle (Ct)
/ see-tee /
In a quantitative PCR, the machine watches a fluorescent signal grow with every cycle. But how do you turn that growing glow into a single number you can compare between samples? You pick a brightness line — a threshold — and ask: at which cycle did the signal first cross it? That cycle number is the threshold cycle, written Ct (or Cq in newer naming).
Early in a qPCR the signal is buried in background noise; as the target multiplies exponentially, fluorescence rises and eventually pokes clearly above the noise. The threshold is a fixed line set just above that noise. The Ct is the (often fractional) cycle at which the curve crosses the threshold. The logic is simple: a sample loaded with target needs few doublings to become detectable, so it crosses early and has a low Ct; a sample with little target needs many doublings, crosses late, and has a high Ct. Because growth is exponential, a difference of one Ct corresponds to about a twofold difference in starting amount, and a difference of about 3.3 cycles to roughly tenfold.
Ct is the number qPCR results are built on — relative gene-expression comparisons, viral-load reports, and the cut-off that decides a clinical 'positive' or 'negative.' But it must be read with care. Ct depends on the assay's efficiency, the threshold setting, and the instrument, so raw Ct values are not portable between labs without calibration, and a very high Ct near the limit of detection is unreliable. Ct measures the amount, not the identity, of what was amplified — confirming you amplified the right thing still requires a probe or a melt analysis.
Comparing a gene's expression in treated versus untreated cells, the treated sample has Ct 22 and the untreated Ct 25. The 3-cycle-lower Ct in the treated sample means it started with roughly 2 to the 3rd — about eight times — more of that mRNA.
Each Ct unit is about a twofold change in starting amount.
A raw Ct number is meaningless on its own — it depends on the threshold, assay efficiency, and machine, and near the detection limit it is noisy, so Ct values should be compared only within a calibrated run, not treated as an absolute concentration.