Every LC–MS lab runs its own column, gradient and instrument. As a result, retention times rarely transfer between labs, and this valuable second line of evidence is usually ignored when annotating unknowns.
This new method changes the game. Instead of predicting absolute retention times, it predicts a Retention Order Index (ROI)—a robust, setup-independent ranking of how compounds elute. The result? Highly accurate RT predictions that actually transfer across different experimental setups with minimal calibration. No training data from your setup required.
We are already working on integrating this feature directly into SIRIUS.
Combining SIRIUS molecular structure annotation with transferable retention time prediction will take compound annotation confidence to a whole new level.
F. Kretschmer et al. Nat Methods (2026) 10.1038/s41592-026-03243-2
