Expert system for computer-assisted annotation of MS/MS spectra

Area of Science:

• Proteomics
• Computational Biology
• Mass Spectrometry

Background:

• Peptide identification in mass spectrometry (MS)-based proteomics relies on analyzing fragment spectra.
• Unassigned peaks in tandem-MS (MS/MS) spectra are common, limiting identification accuracy.
• Manual annotation of these peaks is impractical for large-scale proteomics studies.

Purpose of the Study:

• To develop an Expert System for automated annotation of unassigned peaks in MS/MS spectra.
• To improve the accuracy and completeness of peptide identification in proteomics.
• To provide a tool that assists both novice and expert users in MS-based proteomics data analysis.

Main Methods:

• Developed an Expert System using literature knowledge and high-resolution fragmentation spectra.
• Established a false discovery rate calculation method using random peak insertion for knowledge base optimization.
• Applied the system to high-resolution HCD and CID data.

Main Results:

• The Expert System significantly increases fragment peak annotation coverage in MS/MS spectra.
• For high-resolution HCD data, median intensity coverage increased from 58% to 86%.
• The system's annotation performance exceeds that of human experts, particularly for complex spectra.

Conclusions:

• The developed Expert System effectively annotates unassigned peaks in MS/MS spectra, enhancing peptide identification.
• This computational approach provides expert-level insights, aiding researchers in MS-based proteomics.
• The tool is accessible via MaxQuant and a webserver, supporting broader adoption in the scientific community.