SIMSI-Transfer

SIMSI-Transfer transfers peptide identifications between similar MS2 spectra to reduce missing values and improve identification rates in isobaric stable isotope-labeled proteomic and phosphoproteomic datasets generated with tandem mass tags (TMT).

Key Features:

• Clustering-based identification transfer: Employs a clustering algorithm, extending MaRaCluster, to group similar tandem MS2 spectra across multiple TMT experiments under the assumption that clustered spectra represent identical peptides across batches.
• Reduction of missing values: Transfers peptide identifications from identified to unidentified TMT batches where the same peptide was fragmented but not identified, thereby reducing missing values as sample numbers increase.
• Error control and validation: Validated with masked search engine identification results, recovering over 80% of masked identifications while maintaining an error rate below 1% false discovery rate.
• Enhanced data completeness: Application to six published full proteome and phosphoproteome datasets from the Clinical Proteomic Tumor Analysis Consortium increased identified MS2 spectra with TMT quantifications by 26–45%.
• Increased identification rates: Increased the number of peptides identified across TMT batches by 43–56% and proteins by 13–16%.
• MaRaCluster integration: Builds upon and extends MaRaCluster functionality specifically for TMT data analysis.

Scientific Applications:

• Large-scale TMT proteomics and phosphoproteomics: Improves quantification completeness and identification rates across many TMT-labeled samples in high-throughput studies.
• Cancer proteogenomics (CPTAC datasets): Enhances data completeness and peptide/protein identification in Clinical Proteomic Tumor Analysis Consortium datasets used for cancer research.

Methodology:

Clustering similar MS2 spectra across multiple TMT experiments and transferring peptide identifications from identified to unidentified batches based on spectral similarity, with validation using masked search engine identification results demonstrating >80% recovery and <1% false discovery rate.