Computing exact p-values for a cross-correlation shotgun proteomics score function.

Area of Science:

• Proteomics
• Computational Biology
• Mass Spectrometry

Background:

• Accurate scoring of peptide-spectrum matches is crucial for protein identification in mass spectrometry.
• The SEQUEST XCorr score is a widely used metric, but its interpretation can be affected by spectrum-specific variations.

Purpose of the Study:

• To introduce a novel procedure for computing exact p-values for the SEQUEST XCorr score.
• To enhance the reliability and accuracy of peptide identification in mass spectrometry data analysis.

Main Methods:

• Utilized dynamic programming to efficiently enumerate the score distribution for peptides near the precursor mass.
• Developed a method to compute exact p-values for SEQUEST XCorr scores.
• Integrated the p-value calculation with the Percolator post-processing tool.

Main Results:

• Ranking spectra by p-value significantly reduces score variance attributed to spectrum-specific effects.
• The XCorr p-value approach, combined with Percolator, identified more spectra and peptides compared to Mascot, X!Tandem, Comet, and MS-GF+ at a fixed false discovery rate.
• Demonstrated improved performance across diverse mass spectrometry datasets.

Conclusions:

• Exact p-value computation for SEQUEST XCorr provides a more robust measure for peptide identification.
• This method enhances the statistical rigor of protein mass spectrometry analysis.
• The p-value approach offers superior performance in identifying peptides and spectra compared to existing methods.