Retention time prediction for 653 pesticides on a biphenyl liquid chromatography stationary phase using machine learning

Suspect screening large numbers of analytes by a single LC-MS/MS method has become more widespread in recent years with new advances in high speed data-dependent (DDA) or data-independent (DIA) acquisition methods. The process of molecular identification can however be challenging when it is not possible to measure an authentic standard. Retention time verification (or prediction) is a critical tool in suspect screening. The ability to predict retention times on C18has recently been demonstrated using machine learning tools, but models have not been explored for other reversed-phase media which may offer alternative selectivity to enhance component identification. In this work, the prediction of retention times for a diverse chemical space is considered using artificial neural networks for a biphenyl stationary phase.