Efficient Bayesian protein inference (EPIFANY)

Efficient Bayesian protein inference (EPIFANY) performs probabilistic protein inference from bottom-up proteomics mass spectrometry data using Bayesian networks to resolve ambiguities caused by shared peptides.

Key Features:

• Bayesian Protein Inference: Applies Bayesian network modeling to infer protein presence from peptide-spectrum match (PSM) evidence.
• Shared Peptide Resolution: Addresses ambiguity in protein identification caused by peptides shared among multiple proteins.
• Scalable Large-Scale Analysis: Processes proteomics datasets containing hundreds of thousands of spectra with high computational efficiency.
• Improved Identification Accuracy: Enhances protein identification performance at controlled protein false discovery rates.
• Benchmark Validation: Demonstrates performance on the 2016 iPRG protein inference benchmark dataset.

Scientific Applications:

• Bottom-Up Proteomics Analysis: Infers protein presence from peptide evidence generated by mass spectrometry experiments.
• Protein Identification Studies: Improves detection of true-positive proteins in complex proteomics datasets.
• Proteomics Benchmark Evaluation: Evaluates protein inference methods using benchmark datasets such as the 2016 iPRG protein inference challenge.

Methodology:

EPIFANY performs Bayesian protein inference by applying loopy belief propagation and convolution trees within Bayesian networks to integrate peptide-spectrum match evidence and estimate protein probabilities.