# Generating Dynamic Structures Through Physics‐Based Sampling of Predicted Inter‐Residue Geometries

## TL;DR

This paper introduces trRosettaX2-Dynamics, a method that combines deep learning and physics-based sampling to predict dynamic protein structures and alternative conformations.

## Contribution

The novel contribution is the integration of physics-based sampling with deep learning predictions to model protein dynamics without needing native structural information.

## Key findings

- trRosettaX2-Dynamics effectively predicts alternative conformations and dynamic structures.
- The method was fine-tuned on 7000 dynamic NMR structures, improving its performance.
- Benchmarking shows promising results across datasets focused on conformational and dynamic structures.

## Abstract

Deep learning‐based methods, such as AlphaFold2, have revolutionized the prediction of static protein structures. However, modeling alternative conformations and dynamic structures remains an unsolved problem. Here, we present trRosettaX2‐Dynamics (trX2‐D), an innovative solution building on our CASP15 and CASP16 winning method, trRosettaX2. trX2‐D tackles this challenge by employing physics‐based iterative sampling of trRosettaX2's predicted inter‐residue geometric distributions. The model underwent pre‐training on high‐resolution X‐ray structures, followed by fine‐tuning on approximately 7000 dynamic NMR structures. This dual training regime significantly bolsters its capacity to predict alternative conformations and dynamic structures. At its core, trX2‐D employs a Transformer‐based neural network to initially predict a set of inter‐residue geometric constraints. These constraints are then iteratively sampled to generate dynamic structures, entirely circumventing the need for prior knowledge of native structural states. Extensive benchmarking across three distinct datasets—two focused on alternative conformations and one on dynamic structures—demonstrates trX2‐D's promising ability to predict alternative conformations and accurately capture structural dynamics. This work highlights the potential of integrating deep learning predictions with physics‐based sampling to advance the field of protein dynamic structure prediction.

While static structure prediction has been revolutionized, modeling protein dynamics remains elusive. trRosettaX2‐Dynamics is presented to address this challenge. This framework leverages a Transformer‐based network to predict inter‐residue geometric constraints, guiding conformation generation via physics‐based iterative sampling. The resulting method effectively explores alternative conformations without requiring prior knowledge of native states.