#37 · Structural bio & protein design
MD-informed models map allosteric couplings in receptors and AAA+ ATPases
The common thread across these three preprints is the use of molecular dynamics to resolve allosteric coupling at residue resolution, followed by a statistical or kinetic layer that ranks or validates the resulting pathways. The AT1R study applies interpretable Bayesian network modeling to MD ensembles, mapping residue-level allosteric couplings that are validated by mutational scanning and that yield a negative allosteric modulator. The AAA+ ATPase study pairs comparative MD with SHAP feature ranking to trace allosteric paths in katanin and ClpB. The acetylcholine receptor study combines kinetic modelling, phi-value and activation energy estimates from single-channel recordings with MD to map parallel gating pathways. What is new across these items is the explicit coupling of simulation-derived pathways to interpretability methods or experimental kinetics that constrain them, rather than MD alone.
Preprints
Interpretable Machine Learning Model of Receptor Dynamics Reveals AT1R Allostery and a Negative Allosteric Modulator
Interpretable Bayesian network modeling of MD ensembles maps residue-level allosteric couplings in AT1R, validated by mutational scanning and yielding a…
Probing Mechanisms of Allosteric Regulation in AAA+ ATPases for Microtubule Severing and Protein Disaggregation
Comparative molecular dynamics plus machine learning (SHAP feature ranking) of katanin and ClpB AAA+ ATPases maps allosteric paths linking nucleotide and…
Allosteric signal initiation and communication in neuromuscular acetylcholine receptors
Combines kinetic modelling, phi-value and activation energy estimation from single-channel recordings, and MD simulations to map parallel gating pathways and…