on 06-19-202003:39 PM - edited on 10-15-202108:19 AM by Closed Account
Siegfried Gessulat1, Tobias Schmidt2, Michael Graber1, Florian Seefried1, Carmen Paschke3, Kai Fritzemeier3, Dave Horn4, Bernard Delanghe3, Daniel Zolg2, Mathias Wilhelm2, Bernhard Kuster2, Martin Frejno1 ASMS 2020 Purpose: Improve the separation of target and decoy identifications in proteomics data sets in order to boost the confidence in search results.
Methods: Several datasets were analysedusing a beta version of Thermo Scientific™ Proteome Discoverer™ 2.5 software with SequestHTand the new Prosit-derived (1) Rescoring node by MSAID.
Results: Deep-learning-based prediction of fragment ion intensities enables the addition of intensity-based scores to identification workflows with SequestHT, which increase the confidence in search results.
1msAid GmbH, Garching, Germany; 2Technical University of Munich, Freising, Germany; 3Thermo Fisher Scientific (Bremen) GmbH, Bremen, Germany; 4Thermo Fisher, San Jose, CA