Reflections from the “When Data Science Meets Molecular Dynamics” workshop

The When data science meets molecular dynamics workshop, held from October 15–17, 2025, brought together 47 onsite and 95 online participants to explore how data science can enhance molecular dynamics (MD) simulations and accelerate discoveries in chemistry, biology, and biomedicine.

Over three days, participants engaged with cutting-edge research spanning computational microscopy, protein-membrane interactions, drug design, enhanced sampling methods, and AI-driven analyses. Highlights included talks on large-scale MD data analysis, FAIR-compliant simulation repositories, and deep-learning approaches for predicting protein-ligand interactions and macromolecular dynamics. The workshop also focused on challenges facing the MD community, such as ensuring reproducibility, developing standards for simulation data and metadata, and building sustainable infrastructures for sharing and analyzing trajectories. Discussions emphasized the importance of open, high-quality databases to enable meta-analyses and support machine learning applications. Through lectures, poster sessions, and interactive discussions, participants reflected on the evolving role of MD as a data-rich science and explored strategies to transition from traditional, single-lab simulations to a collaborative, data-driven paradigm. The workshop concluded with a shared vision for a more FAIR, reproducible, and AI-ready MD community.