Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

BP: Fachverband Biologische Physik

BP 39: Focus Session: Theoretical Modeling and Simulation of Biomolecular Condensates III (joint session CPP/BP)

BP 39.1: Topical Talk

Freitag, 13. März 2026, 09:30–10:00, ZEU/0260

Data-driven modelling of phase-separating intrinsically disordered regions — •Giulio Tesei^1,2, Fatima Kamal Zaidi³, Shanlong Li⁴, Julian O. Streit¹, Jianhan Chen⁴, Tanja Mittag³, and Kresten Lindorff-Larsen¹ — ¹Department of Biology, University of Copenhagen, Copenhagen, Denmark — ²Department of Biomedical Science, Malmö University, Malmö, Sweden — ³Department of Structural Biology, St. Jude Children's Research Hospital, Memphis, U.S.A. — ⁴Department of Chemistry, University of Massachusetts, Amherst, U.S.A.

Intrinsically disordered regions (IDRs) constitute about one third of the human proteome and play important roles in biological processes. While lacking well-defined 3D structures, IDRs adopt heterogeneous ensembles influenced by multivalent interactions; these same interactions can promote phase separation and contribute to the formation of biomolecular condensates. I will first present CALVADOS, an efficient one-bead-per-residue model optimized on experimental data reporting on IDR conformational properties and extensively validated on both single-chain and phase behavior across diverse sequences. I will then describe how we used large sets of CALVADOS simulations to train machine-learning models that accurately predict single-chain compaction and homotypic phase-separation propensity directly from sequence. Finally, I will introduce a hybrid-resolution model with an atomistic backbone representation that matches the accuracy of CALVADOS for global dimensions and phase separation while also capturing local structure and backbone hydrogen bonding.

Keywords: Intrinsically disordered regions; Biomolecular condensates; Force field parameterization; Coarse-grained modeling