Comparative Analysis of Data-Driven Rescoring Platforms for Improved Peptide Identification in HeLa Digest Samples

Area of Science:

• Proteomics and Mass Spectrometry
• Bioinformatics and Computational Biology

Background:

• Mass spectrometry is essential for analyzing biological processes, but many spectra remain unassigned.
• Advances in search engine technology have not fully resolved the issue of unassigned spectra in proteomics.

Purpose of the Study:

• To evaluate the performance of three rescoring platforms: Oktoberfest, MS²Rescore, and inSPIRE.
• To compare rescoring platform effectiveness against original search engine results using MaxQuant output.

Main Methods:

• Evaluation of Oktoberfest, MS²Rescore, and inSPIRE platforms using MaxQuant search results.
• Analysis of peptide and PSM (Peptide-Spectrum Match) identification rates.
• Assessment of factors influencing platform performance, including PTM compatibility and computational time.

Main Results:

• Rescoring platforms increased peptide identifications by 40%-53% and PSMs by 64%-67%.
• Up to 75% of lost peptides exhibited posttranslational modifications (PTMs), highlighting limitations in PTM processing.
• inSPIRE excelled in peptide and unique peptide identifications; MS²Rescore performed better for PSMs at higher FDRs.

Conclusions:

• Rescoring platforms significantly improve proteomic data analysis over original search results.
• Platform performance varies due to search engine features, ion series prediction, retention time prediction, and PTM compatibility.
• Integration of rescoring platforms is necessary but requires addressing computational costs and optimization challenges.