Comparative evaluation of tandem MS search algorithms using a target-decoy search strategy

Area of Science:

• Proteomics
• Mass Spectrometry
• Bioinformatics

Background:

• Peptide identification from tandem mass spectra is crucial for proteomics.
• Inconsistent application and understanding of search algorithms hinder reliable results.
• Evaluating and interpreting algorithm outputs requires standardized methods.

Purpose of the Study:

• To compare four tandem mass spectrometry peptide identification search algorithms: Mascot, Open Mass Spectrometry Search Algorithm (OMSSA), Sequest, and X! Tandem.
• To assess the impact of scoring methodologies and a target-decoy approach on algorithm performance.
• To propose an alternative method for determining optimal cutoff thresholds.

Main Methods:

• Conducted a survey of four peptide identification algorithms using identical input data, search parameters, and sequence libraries.
• Compared algorithm outputs based on common scoring methodologies.
• Utilized a target-decoy approach for sequence library searching analysis.

Main Results:

• Little difference in algorithm output was observed when consistent scoring procedures were applied.
• Certain commonly used scoring procedures were found to potentially lead to excessive false discovery rates.
• An alternative method for determining optimal cutoff thresholds was proposed.

Conclusions:

• Algorithm choice has minimal impact on peptide identification results if scoring is consistent.
• Careful selection and application of scoring methodologies are essential to control false discovery rates.
• The proposed alternative method offers a more reliable approach for setting optimal cutoff thresholds in proteomics data analysis.