Welcome to MHCXGraph’s documentation!¶
Welcome to the MHCXGraph documentation. This page is designed to help you get started with our graph-based structural similarity detection framework for biomolecular complexes.
MHCXGraph is an open-source Python package for detecting structurally conserved surface regions across protein structures, with a particular focus on peptide–MHC (pMHC) complexes involved in T cell receptor (TCR) recognition. The method represents biomolecular surfaces as graphs of solvent-exposed residues and identifies similarities through a triad-based geometric and physicochemical comparison framework.
The approach combines discretized representations of local residue environments with a restricted Cartesian product across multiple structures, enabling the detection of conserved spatial patterns that may underlie TCR cross-reactivity. The package also provides interactive visualization interfaces for exploring the resulting association graphs and structural correspondences.
About