MSblender: A probabilistic approach for integrating peptide identifications from multiple database search engines

Area of Science:

• Proteomics
• Mass Spectrometry
• Bioinformatics

Background:

• Shotgun proteomics is vital for protein identification but has limited sensitivity in complex samples.
• Integrating peptide identifications from multiple search engines can enhance protein identification.
• Current integration methods struggle with reliable error rate control.

Purpose of the Study:

• To develop a statistically coherent method for integrating peptide identifications from multiple search engines.
• To improve the sensitivity and accuracy of protein identification and quantification in shotgun proteomics.
• To enable reliable estimation of false discovery rates in integrative proteomic analysis.

Main Methods:

• Developed MSblender, a novel statistical method for integrative proteomic data analysis.
• MSblender converts raw search engine scores into probability scores for peptide-spectrum matches (PSMs).
• The method accounts for correlations between search scores and reliably estimates false discovery rates.

Main Results:

• MSblender identifies more PSMs than individual search engines at the same false discovery rate.
• Increased PSM identifications lead to improved protein spectral counts and quantification.
• Enhanced quantification improves sensitivity in differential expression analyses.

Conclusions:

• MSblender offers a statistically sound approach for integrating proteomic search engine results.
• The method enhances protein identification and quantification, particularly in complex biological samples.
• MSblender advances the field of quantitative proteomics and differential expression analysis.