PIA - Protein Inference Algorithms

PIA - Protein Inference Algorithms performs protein inference by integrating peptide-spectrum matches (PSMs) from multiple database search engines to produce consistent protein identifications for LC-MS/MS proteomics.

Key Features:

• PSM integration and consolidation: Integrates peptide-spectrum matches (PSMs) from multiple database search engines and consolidates them into unified evidence for protein inference.
• Multi-engine and mzIdentML support: Supports the majority of established search engines and reads mzIdentML-formatted results for spectrum identification.
• PSI standard formats: Supports PSI standard file formats used for spectrum identification and protein inference.
• Protein inference algorithms: Implements several established protein inference algorithms to infer proteins from peptide evidence and to achieve consistent false discovery rate (FDR) thresholds across combined results.
• Workflow integration: Provides nodes for integration into workflow environments such as the KNIME Analytics Platform.
• Benchmarking and isoform detection: Demonstrated on benchmark datasets to increase protein identification, including isoform detection, compared with single-engine inferences.

Scientific Applications:

• High-throughput LC-MS/MS proteomics: Inferring proteins from peptide-centric LC-MS/MS datasets where direct protein measurement is limited.
• Peptide-based identification and quantification: Converting peptide identification and quantification results into reliable protein-level identifications for biological and disease-focused studies.

Methodology:

PIA integrates PSMs from multiple database search engines (including mzIdentML/PSI formats) and applies implemented protein inference algorithms to derive protein-level identifications and report isoform-relevant assignments.