Monday Webinar: Statistical inference of omics data with variable-selection

Liquid chromatography with tandem mass spectrometry (LC-MS/MS) is one of the most effective analytical techniques to characterize biological samples. The identification of molecules depends on the LC-MS/MS instrumentation and experimental settings and it is likely that a molecule detected in a metabolomics trial may not be present in the reference library.

Computational approaches can be used to enhance spectral matching. In particular, chemometrics and deep-learning-based approaches have recently been applied to directly predict molecular structures from large databases of MS spectra. Molecular structures can be numerically represented through in-silico fingerprints, which are binary vectors that encode features of molecules. Prediction of molecular fingerprints starting from the LC-MS/MS spectra would consequently assist the match of target compounds, which would benefit from the increased dimension of fingerprint databases.