Experiments in searching small proteins in unannotated large eukaryotic genomes

Area of Science:

• Proteomics
• Genomics
• Bioinformatics

Background:

• Mass spectrometry data is increasingly used for direct genome sequence searching.
• This approach can correct and complement existing genome annotations.
• Searching large eukaryotic genomes with peptide tandem mass spectra presents practical difficulties, especially for small proteins (<40 kDa).

Purpose of the Study:

• To explore the challenges of automatically identifying peptides that span across exon/intron boundaries using experimental data.
• To assess the impact of mass spectrometry data on genome annotation accuracy.
• To demonstrate the utility of proteomics data in refining predicted gene structures.

Main Methods:

• Directly searching large eukaryotic genome sequences using peptide ion trap tandem mass spectra.
• Analyzing peptides that span across exon/intron boundaries.
• Evaluating search results against Swiss-Prot for comparison.
• Assessing the effect of parent mass accuracy on peptide identification rates.

Main Results:

• Approximately 30% of peptides were missed in a human genome search compared to a Swiss-Prot search.
• This peptide miss rate was significantly reduced with improved parent mass accuracy.
• Several examples of predicted gene structures were identified that could benefit from proteomics data, particularly peptides spanning exon/intron boundaries.

Conclusions:

• Directly searching genome sequences with mass spectrometry data is a valuable approach for improving genome annotations.
• Identifying peptides spanning exon/intron boundaries is a key challenge but crucial for accurate gene structure prediction.
• Enhanced mass accuracy in proteomics data significantly improves the identification of relevant peptides, aiding in the refinement of gene models.