MyriMatch

MyriMatch applies a multivariate hypergeometric statistical scoring model to MS/MS data to improve peptide identification in shotgun proteomics by emphasizing the intensity of matched peaks.

Key Features:

• Statistical scoring model: Uses a multivariate hypergeometric distribution to score peptide-spectrum matches with emphasis on intense matched peaks.
• Intensity-based stratification: Stratifies peaks into multiple intensity classes to enhance discrimination between correct identifications and random matches.
• Random-chance probability: Calculates the probability that the best match for each spectrum has occurred by chance to separate true positives from false positives.
• Performance evaluation and comparison: Demonstrated superior discrimination compared to Sequest and X!Tandem across datasets from three laboratories using various ion trap instruments, and shows performance decline with minimal peak filtering.
• Scalability: Scales efficiently with tandem mass spectra from high-resolution mass analyzers.

Scientific Applications:

• Peptide identification in shotgun proteomics: Improves reliability of peptide assignments in shotgun proteomics by incorporating peak intensity into scoring.
• MS/MS data analysis across instrument types: Applicable to MS/MS datasets from various ion trap instruments and high-resolution mass analyzers.
• Search engine benchmarking: Serves as a basis for comparative evaluation of database search algorithms such as Sequest and X!Tandem using multi-laboratory datasets.

Methodology:

Employs a multivariate hypergeometric distribution-based scoring model, stratifies peaks into multiple intensity classes, computes per-spectrum random-chance probabilities for best matches, and was evaluated on datasets from three laboratories using various ion trap instruments with comparisons to Sequest and X!Tandem; performance is sensitive to the extent of peak filtering and the method scales to high-resolution mass analyzer spectra.