20/01/2026
We are pleased to announce the publication of our latest work in the Journal of Chemical Theory and Computation!
Determining the Optimal Structural Resolution of Proteins through an Information-Theoretic Analysis of Their Conformational Ensemble, by Margherita Mele, Raffaele Fiorentini, Thomas Tarenzi, Giovanni Mattiotti and Raffaello Potestio
https://pubs.acs.org/doi/10.1021/acs.jctc.5c01773
How much detail do we really need to understand protein dynamics?
In protein modeling and analysis, more detail isn’t always better. While atomistic simulations capture extremely fine features of the molecule's structural fluctuations, much of that information turns out to be redundant when we’re trying to understand the key motions and conformational states that matter for function.
In our new work, we introduce an unsupervised, information-theoretic framework that answers a simple but fundamental question: what is the minimal structural resolution needed to retain the essential information about a protein’s behavior?
By systematically removing atomic degrees of freedom and measuring how much configurational information is preserved, we find that:
- The optimal level of detail scales linearly with protein size.
- On average, about four heavy atoms per residue are sufficient to capture the relevant conformational landscape.
- Remarkably, this matches the resolution of widely used coarse-grained models like MARTINI and SIRAH.
- Proteins dominated by large, collective motions require even less detail.
Overall, this work provides a general way to identify the simplest model that is still maximally informative, offering a new perspective on the relationship between protein structure, dynamics, and function - and a guide for building efficient multiscale models.
