FIT: Tag based method for fusion proteins identification

There is increased interest in the identification and analysis of gene fusions and chimeric RNA transcripts. While most recent efforts focused on the analysis of genomic and transcriptomic data, identification of novel peptides corresponding to such events in mass spectrometry-based proteomic datasets would provide complementary, protein-level evidence. The process of identifying fusion proteins from mass spectrometry data is inherently difficult because such events are rare. It is also complicated due to large amount of spectra collected and the explosion in the number of candidate peptide sequences that need to be considered, which makes exhaustive search for all possible fusion partner proteins impractical. In this work, we present a sequence tag based fusion protein identification algorithm, FIT, that combines the virtue of both de novo sequence tag retrieval and peptide-spectrum matching for identification of fusion proteins. Results on simulated datasets show high sensitivity and low false positive rates for fusion protein identification by the FIT algorithm.