Peptide identification is important in bottom-up proteomics.
Post-translational
modifications (PTMs) are crucial in regulating cellular activities.
Many database search methods have been developed to identify peptides
with PTMs and characterize the PTM patterns. However, the PTMs on
peptides hinder the peptide identification rate and the PTM characterization
precision, especially for peptides with multiple PTMs. To address
this issue, we present a sensitive open search engine, PIPI2, with
much better performance on peptides with multiple PTMs than other
methods. With a greedy approach, we simplify the PTM characterization
problem into a linear one, which enables characterizing multiple PTMs
on one peptide. On the simulation data sets with up to four PTMs per
peptide, PIPI2 identified over 90% of the spectra, at least 56% more
than five other competitors. PIPI2 also characterized these PTM patterns
with the highest precision of 77%, demonstrating a significant advantage
in handling peptides with multiple PTMs. In the real applications,
PIPI2 identified 30% to 88% more peptides with PTMs than its competitors.