LineageFilter:\nImproved Proteotyping of Complex Samples\nUsing Metaproteomics and Machine Learning

Metaproteomics is\na powerful tool to characterize how\nmicrobiota\nfunction by analyzing their proteic content by tandem mass spectrometry.\nGiven the complexity of these samples, accurately assessing their\ntaxonomical composition without prior information based solely on\npeptide sequences remains a challenge. Here, we present LineageFilter,\na new python-based AI software for refined proteotyping of complex\nsamples using metaproteomics interpreted data and machine learning.\nGiven a tentative list of taxa, their abundances, and the scores associated\nwith their identified peptides, LineageFilter computes a comprehensive\nset of features for each identified taxon at all taxonomical ranks.\nIts machine-learning model then assesses the likelihood of each taxon’s\npresence based on these features, enabling improved proteotyping and\nsample-specific database construction.