Conformational dynamics are essential for a wide range of biological processes including the folding of proteins into their functional conformations, toggling between alternative conformations in signaling, self-assembly and aggregation, and association/dissociation of small molecules and other binding partners.

Markov state models (MSMs) are proving to be invaluable tools for interrogating these processes and providing the mechanistic understanding required to control them with mutations, small molecules, or other perturbations. In recent years, Machine learning (ML)¹ has revolutionized the field of biomolecular dynamics, leading to the development of new algorithms that can enhance the modeling of biomolecular dynamics. This surge of interest has led to the development of numerous physics-² and ML-based algorithms^{3, 4} to facilitate modeling of conformational dynamics (e.g., to perform adaptive sampling, conduct dimensionality reduction, and identify metastable states). In addition, emerging techniques⁵ that encode non-Markovian dynamics via the Generalized Master Equation (GME)⁶ or Generalized Langevin Equation (GLE)⁷ have also been developed to improve the performance of MSMs for studying biomolecular dynamics.

In this workshop, our aim is to create a platform for active scientists in the field to discuss recent advancements in machine learning algorithms for studying conformational dynamics, particularly using MSM and related theoretical frameworks. Some specific topics of interest to this workshop involve, but are not limited to:

• Advancements in developing physics- or ML-based methods to identify collective variables or committer functions for biomolecular conformational changes.
• Generative-AI methods to model and sample biomolecular conformational changes in the latent space.
• New approaches to perform adaptive sampling for the facilitating of the MSM construction.
• Recent development of non-Markovian dynamic models for biomolecular dynamics based on the GME or GLE framework.
• Development of re-weighing methods for both thermodynamics and dynamics for biomolecular conformational changes.
• Application of MSMs and related techniques (e.g., Milestoning and Weighted Ensemble simulations) to elucidate molecular mechanisms underlying important biological processes, and facilitate drug design (e.g., identify cryptic pockets⁸).

Our workshop will draw multidisciplinary expertise spanning chemistry, physics, applied mathematics, and computer science from around the globe. This platform will enable discussions on recent methodological developments, new mechanistic insights, and innovative hypotheses concerning the conformational dynamics underpinning diverse biological and chemical processes. In the era of machine learning, we anticipate that the rapid evolution of MSM and related methodologies will broaden their application in studying conformational dynamics. We aim for our workshop to stimulate valuable interactions among leading scientists in this dynamic field, fostering new collaborations and advancements.