A statistical model for identifying proteins by tandem mass spectrometry

Area of Science:

• Proteomics and Bioinformatics
• Statistical Modeling in Mass Spectrometry

Background:

• Accurate protein identification from mass spectrometry (MS/MS) data is crucial for biological research.
• Existing methods face challenges in handling peptides assigned to multiple proteins and quantifying identification reliability.

Purpose of the Study:

• To develop a statistical model for calculating protein presence probabilities in biological samples.
• To establish a reliable method for filtering large-scale proteomics datasets with controlled false positive rates.

Main Methods:

• A statistical model was developed to compute protein presence probabilities based on peptide assignments to MS/MS spectra.
• Peptides mapping to multiple proteins were apportioned, and the expectation-maximization algorithm identified a minimal protein list.
• The model was validated using purified proteins and complex biological samples (H. influenzae, Halobacterium).

Main Results:

• The model accurately computes protein presence probabilities, demonstrating high power in distinguishing correct from incorrect identifications.
• Demonstrated ability to filter large-scale proteomics data with predictable sensitivity and false positive error rates.
• The approach is fast, consistent, and transparent, facilitating reproducible large-scale proteomics studies.

Conclusions:

• The presented statistical model offers an accurate and powerful method for protein identification in proteomics.
• This approach provides a standardized framework for publishing and comparing large-scale proteomics data.
• Enables reliable filtering of proteomics datasets, improving the quality and interpretability of results.