Annotated Protein Database Using Known Cleavage Sites for Rapid Detection of Secreted Proteins

Liquid
chromatography
tandem mass spectrometry
(LC–MS/MS) analysis of secreted proteins has contributed to
our understanding of human disease and physiology but is limited by
its need for accurate protein database annotation. Common assumptions
used in proteomics of perfect protease specificity are inaccurate
for secreted proteins, which are cleaved by numerous endogenous proteases.
Here, we describe the generation of an optimized protein database
that divides proteins into their individual biological chains and
peptides to allow fast identification of semi-tryptic peptides from
secreted proteins using fully tryptic searches. We applied this biologically
annotated database to previously published human plasma proteome data
sets containing either DIA or DDA data, using Spectronaut, DIA-NN,
MaxDIA, and MaxQuant. Using our annotated database, we greatly reduced
search times while achieving similar protein and peptide identifications
compared to that obtained from standard approaches using semi-tryptic
searches. Furthermore, our database enables the identification of
biologically relevant semi-tryptic peptides using data analysis packages
that are not capable of semi-tryptic searches. Together, these findings
demonstrate that our annotated database is more capable than currently
available databases for secreted protein analysis and is particularly
useful for large-scale plasma proteome analysis.