SpecOMS, an open modification search approach challenging high-throughput single amino acid variations identification
Résumé
Proteomics studies are based on the identification of proteins. Although biochemistry assays and
bioinformatics strategies have made significant advances in this field, they are still unaware of a
comprehensive solution. Generally, the protein identification methodology consists of spectrum
similarity search by comparing an experimentally acquired fragmentation spectrum against a
theoretical spectrum derived from a protein sequence database. Additionally, to process a large universe
of pairwise comparisons, most search engines start by selecting theoretical peptides based on their
mass. Since the mass modifications must be declared in advance, it represents a challenge to identify
unknown post-translation modifications (PTMs) and single nucleotide variants (SAVs).
The Open Modification Search algorithms (OMS) use fast calculation methods to identify proteins
without holding the mass precursor. Also, the expected mass changes do not need to be declared in
advance. In this approach, the first step is to identify all peptides, then the second step is to interpret the
mass delta between the precursor ion and the theoretical peptide in terms of PTM or SAV.
In this work, we used the OMS algorithm called SpecOMS with the aim to evaluate its performance for
SAVs identification. We compare our results with those obtained from X!Tandem, which uses a classic
approach to identify peptides. A gold standard dataset was used: four corn lineages (Zea mays - B73,
F2, EA1192, and MBS847), each one already sequenced. In order to evaluate the ability of SpecOMS
in identifying SAVs, we performed
self and cross-interrogations over the databases. The false
discovery rate (FDR) estimation was implemented to qualify the peptide spectrum match (PSM).
Finally, we will present our results in terms of the FDR implementation, refinement parameters and the
performance of SpecOMS to identify SAVs.