Return
Macroscopic joint-based analysis of protein ensembles generated by molecular dynamics simulation
DOI:10.1007/s12257-026-00319-w.png)
Abstract
En 中文
Molecular dynamics (MD) simulation is a powerful tool for generating protein conformational ensembles, yet the resulting high-dimensional data present significant challenges for traditional atomic-level structural analysis. To address this, we present a complementary and scalable method: the joint-based descriptor (JBD), a macroscopic geometric framework that encodes protein conformation through a minimal set of dihedral angles measured at the joints of secondary structures. Human vitamin K reductase (hVKOR), an essential three-transmembrane (TM) anticoagulant membrane enzyme, served as a model system. We analyzed approximately 2,250 structural models generated from a 200-ns MD simulation. Using three specific joint dihedral angles (Ω1, λ1, and Ω2), the JBD mapped the dynamic topology of the hVKOR ensemble. The analysis revealed tightly restricted helical arrangements and identified a discrete and characteristic λ1 range (–80 to –50) as a characteristic topological signature for hVKOR. We further computed a Jscore metric to quantify conformational distances between hVKOR and other 3TM protein families, enabling hierarchical clustering of closest relatives while separating distant topologies. Finally, we implemented a conformational mimicry bias analysis based on directional Jtotal differences to resolve non-reciprocal structural relationships across the 3TM proteome, revealing asymmetric conformational mimicry in which specific families exhibit a directional bias to adopt the hVKOR signature. Our results establish Jscore quantification as a scalable approach for model curation and proteome-wide topology mapping, bridging static AI-predicted structures and dynamic conformational landscapes of proteins.
Keywords:
Macroscopic descriptor
Conformational clustering
Jscore
Molecular dynamics
Human vitamin K reductase
Topological signatures
Journal
IF:
3
Papers:
156
Citations:
3.2K

