Peptide retention prediction algorithm and its application in proteomics

Most of the proteomics analysis methods based on tandem mass spectrometry rely on the matching scoring of actual spectrum and theoretical spectrum, the interference of a large number of co-eluting peptides could cause error in the identification and quantification of peptides and proteins. Peptide retention time prediction, as a auxiliary and verification index of the peptide, can transition the chromatographic behavior into stable independent time attributes, and improve the accuracy of the peptide identification. Prediction of peptide chromatographic retention in complex systems is also of great significance for optimizing proteomics determination conditions and improving the detection rate and repeatability of mass spectrometry data in data-independent acquisition. This review focused on the chromatographic retention prediction method of unmodified peptides and modified peptides, summarizes the content, characteristics and limitations of four types of peptide retention time prediction methods based on standardized indexes, peptide molecular models, amino acid residue parameters, and machine learning, as well as their applications in proteomics, with a prospect of their future.