Using Cluster-Unique Sequences in Proteomes to Enhance Confidence in Identifications from Shotgun Proteomics Studies of Organisms with Complex Genomes

Shotgun proteomics approaches using tandem mass spectrometry (MS/MS) identify proteins by sequencing their constituent peptides and informatically mapping them back into proteins. Computationally, this is not trivial for higher eukaryotes, which often suffer from a “protein inference problem” attributable to the high degree of redundancy in the genome, and therefore proteome of the organism. In this scenario, peptides are likely to map to multiple proteins and, consequently, increase ambiguity within protein identification reports.

In a recent study of Populus, clustering the proteome by 90% sequence similarity was demonstrated as an effective method to avoid under- and over-identification of proteins. This approach collapsed analytically-indistinguishable proteins into single representative units, reannotating the proteome’s 45,778 proteins into 35,854 protein groups. Whereas 47.5% of the unclustered proteome contained at least one unique identifier, 75% of the protein groups were distinguishable after clustering. Cluster-unique peptide sequences that unambiguously identify protein groups simultaneously simplify protein counts and confer additional confidence in the biological functions represented.

While a 90% clustering threshold significantly improved the number of distinguishable proteins in the Populus proteome, it has not been established whether other highly complex organisms would benefit from the same threshold. In this study, we systematically investigate how proteome sizes, complexities, and redundancies influence the level of ambiguity for a particular organism. More specifically, we propose choosing appropriate clustering thresholds based on a universal minimum percentage of cluster-unique peptide sequences. This process not only mitigates analytical ambiguity but preserves biological relevance for protein identification across a range of high-complexity organisms.