AX-R11 — Retention time outside compound window

Retention time is the cheapest orthogonal check on identity available, and it is the one most often discarded. A compound's position on a gradient is a property of the compound; if the peak is not where it should be, either the method was not what was described or the analyte is not what was named.

The rule normalises for gradient length before comparing, because a 15-minute gradient and a 35-minute gradient place the same compound at different absolute times. Normalisation is linear in gradient duration and is stated in full on /methods/retention-normalisation/. It is an approximation and it works better for mid-gradient elution than for very early or very late peaks, which is why the short bioregulator class and the most hydrophobic incretins are the two groups where this rule produces the most false positives.

Two thirds of instances resolve to an undocumented method difference: a different column chemistry, a different temperature, a different ion-pairing additive. Those are recorded as method notes rather than as material findings. The remaining third are cases where the identity question is genuine, and in those the index looks for a mass to resolve it.

Where a compound has a known near-neighbour — survodutide and tirzepatide, hexarelin and GHRP-6, bremelanotide and melanotan II — a retention flag plus a mass within tolerance of the neighbour is escalated.