Characterization of Proteoforms with Unknown Post-translational Modifications Using the MIScore

Various
proteoforms may be generated from a single gene due to
primary structure alterations (PSAs) such as genetic variations, alternative
splicing, and post-translational modifications (PTMs). Top-down mass
spectrometry is capable of analyzing intact proteins and identifying
patterns of multiple PSAs, making it the method of choice for studying
complex proteoforms. In top-down proteomics, proteoform identification
is often performed by searching tandem mass spectra against a protein
sequence database that contains only one reference protein sequence
for each gene or transcript variant in a proteome. Because of the
incompleteness of the protein database, an identified proteoform may
contain unknown PSAs compared with the reference sequence. Proteoform
characterization is to identify and localize PSAs in a proteoform.
Although many software tools have been proposed for proteoform identification
by top-down mass spectrometry, the characterization of proteoforms
in identified proteoform–spectrum matches still relies mainly
on manual annotation. We propose to use the Modification Identification
Score (MIScore), which is based on Bayesian models, to automatically
identify and localize PTMs in proteoforms. Experiments showed that
the MIScore is accurate in identifying and localizing one or two modifications.