Proteomic Validation of Transcript Isoforms, Including Those Assembled from RNA-Seq Data

Human
proteome analysis now requires an understanding of protein
isoforms. We recently published the PG Nexus pipeline, which facilitates
high confidence validation of exons and splice junctions by integrating
genomics and proteomics data. Here we comprehensively explore how
RNA-seq transcriptomics data, and proteomic analysis of the same sample,
can identify protein isoforms. RNA-seq data from human mesenchymal
(hMSC) stem cells were analyzed with our new TranscriptCoder tool
to generate a database of protein isoform sequences. MS/MS data from
matching hMSC samples were then matched against the TranscriptCoder-derived
database, along with Ensembl and the neXtProt database. Querying the
TranscriptCoder-derived or Ensembl database could unambiguously identify
∼450 protein isoforms, with isoform-specific proteotypic peptides,
including candidate hMSC-specific isoforms for the genes DPYSL2 and
FXR1. Where isoform-specific peptides did not exist, groups of nonisoform-specific
proteotypic peptides could specifically identify many isoforms. In
both the above cases, isoforms will be detectable with targeted MS/MS
assays. Unfortunately, our analysis also revealed that some isoforms
will be difficult to identify unambiguously as they do not have peptides
that are sufficiently distinguishing. We covisualize mRNA isoforms
and peptides in a genome browser to illustrate the above situations.
Mass spectrometry data is available via ProteomeXchange (PXD001449).