Progress in deep Markov State Modeling: Coarse graining and experimental data restraints

TL;DR

This paper advances deep Markov State Modeling by integrating physical constraints, experimental data, hierarchical layers, and attention mechanisms to better analyze complex biophysical systems like proteins.

Contribution

It introduces new neural network components and methods to incorporate experimental data and hierarchical modeling into deep Markov State Models for biophysical applications.

Findings

Improved modeling of protein dynamics with experimental data integration

Hierarchical neural network layer enhances detail-level analysis

Attention mechanism identifies key residues for state classification

Abstract

Recent advances in deep learning frameworks have established valuable tools for analyzing the long-timescale behavior of complex systems such as proteins. Especially the inclusion of physical constraints, e.g. time-reversibility, was a crucial step to make the methods applicable to biophysical systems. Furthermore, we advance the method by incorporating experimental observables into the model estimation showing that biases in simulation data can be compensated for. We further develop a new neural network layer in order to build an hierarchical model allowing for different level of details to be studied. Finally, we propose an attention mechanism which highlights important residues for the classification into different states. We demonstrate the new methodology on an ultralong molecular dynamics simulation of the Villin headpiece miniprotein.