BioEmu‑1: Protein Modeling at GPU Speed - Deep‑learning model capable of generating thousands of protein structures per hour on a single GPU
Insight in Hours, Not Years - Emulates protein structural ensembles, delivering insights in hours instead of years of traditional molecular dynamics simulation
Cracking the Protein Motion Puzzle - Captures dynamic motions like cryptic pocket formation, local unfolding, and domain rearrangements
Thermodynamic Precision within 1 kcal/mol - Predicts relative free energies with ~1 kcal/mol accuracy, aligned with millisecond‑scale MD and experimental data
Fusion of Simulation and Stability Data – Integrates over 200 ms of molecular dynamics, static protein structures, and experimental stability data via novel training algorithms
From Simulations to Scalable Discovery - Provides mechanistic insight by jointly modeling structural ensembles and thermodynamic properties.