PrInCE

PrInCE infers protein-protein interactions from co-fractionation mass spectrometry (CF-MS) co-elution data using co-elution patterns and a naive Bayes classifier to reconstruct interaction networks.

Key Features:

• Machine Learning Approach: Uses a naive Bayes classifier to process co-fractionation mass spectrometry (CF-MS) co-elution chromatogram profiles and infer protein-protein interactions.
• Enhanced Performance: An R implementation replaced the original Matlab scripts, improving runtime efficiency and reducing memory requirements.
• Detection of Co-Eluting Protein Complexes: Identifies protein complexes based on shared co-elution profiles.
• Differential Network Analysis: Compares interaction networks across conditions to detect dynamic changes in protein interactions.
• Bioconductor Compatibility: Implemented as an R/Bioconductor package and compatible with Bioconductor classes.

Scientific Applications:

• Protein-protein interaction network reconstruction: Constructs PPI networks from CF-MS co-fractionation data to map interaction landscapes.
• Differential interaction analysis: Reveals how protein interactions change under different experimental conditions or disease states.
• Biomarker identification: Supports identification of condition-specific interaction patterns that may serve as biomarkers.
• Therapeutic target exploration: Facilitates exploration of therapeutic targets by highlighting interaction partners and complex composition.

Methodology:

PrInCE applies a naive Bayes classifier trained on dataset-derived features to analyze co-elution chromatogram profiles from co-fractionation mass spectrometry (CF-MS) and predict protein-protein interactions; it is implemented as an R/Bioconductor package and was ported from Matlab to improve runtime and memory usage.