ReX

ReX infers residue-level hydrogen deuterium exchange (HDX) patterns from HDX-MS data by recasting the analysis as a multiple change-point detection problem within a Bayesian non-parametric framework.

Key Features:

• Residue-level inference: ReX leverages peptide overlap, temporal components of the data, and correlation along the sequence dimension to infer residue-level deuterium uptake patterns.
• Bayesian non-parametric modeling: The framework performs parameter number inference, differential HDX confidence assessments, and uncertainty estimation for temporal kinetics.
• Statistical significance testing: ReX infers statistical significance for individual residues by treating HDX-MS as a multiple change-point problem.
• Resolution metrics: The method provides global and local resolution metrics that align HDX-MS with reporting standards of other structural methods.
• Benchmarked performance: Validation using a three-way proteolytic digestion experiment demonstrates superior performance at predicting unseen HDX data compared to existing methods.

Scientific Applications:

• Protein conformational dynamics: ReX enables exploration of conformational dynamics of proteins and protein complexes in solution at residue-level resolution.
• Differential HDX analysis: The method quantifies differential flexibility of protein domains in the presence of ligands, as demonstrated for BRD4's two Bromodomains with I-BET151.
• Ligand mode of action characterization: ReX reveals distinct residue-level HDX signatures for ligands with varied functional outcomes, as shown for a panel of seventeen small molecules on LXRα.

Methodology:

ReX recasts HDX-MS analysis as a multiple change-point detection problem and fits the model in a Bayesian non-parametric framework.