ASV-ID, a Proteogenomic Workflow To Predict Candidate Protein Isoforms on the Basis of Transcript Evidence

One
of the goals of the Chromosome-Centric Human Proteome Project
(C-HPP) is to map and characterize the functions of protein isoforms
produced by alternative splicing of genes. However, identifying alternative
splice variants (ASVs) via mass spectrometry remains a major challenge,
because ASVs usually contain highly homologous peptide sequences.
A routine protein sequence analysis suggests that more than half of
the investigated proteins do not generate two or more uniquely mapping
peptides that would enable their isoforms to be distinguished. Here,
we develop a new proteogenomics method, named “ASV-ID”
(alternative splicing variants identification), which enables identification
of ASVs by using a cell type-specific protein sequence database that
is supported by RNA-Seq data. Using this workflow, we identify 1935
distinct proteins under highly stringent conditions. In fact, transcript
evidence on these 841 proteins helps us distinguish them from other
isoforms, despite the fact that these proteins are not predicted to
make 2 or more uniquely mapping peptides. We also demonstrate that
ASV-ID enables detection of 19 differently expressed isoforms present
in several cell lines. Thus, a new workflow using ASV-ID has the potential
to map yet-to-be-identified difficult protein isoforms in a simple
and robust way.