Molecular Latent Space Simulators

 · Invited

Abstract

We have developed molecular latent space simulators (LSS) to learn highly efficient and accurate surrogate models of molecular dynamics (MD) by stacking three specialized deep learning networks to (i) encode a molecular system into a slow latent space, (ii) propagate dynamics in this latent space, and (iii) generatively decode a synthetic molecular trajectory. The trained LSS generates novel ultra-long molecular trajectories at six orders of magnitude lower cost than MD enabling resolution of rare thermodynamic states and kinetic transitions with arbitrarily low statistical uncertainties. In an application to Trp-cage, we generate millisecond trajectories in just minutes of wall clock time and demonstrate excellent agreement with the MD structure, thermodynamics, and kinetics.

H. Sidky, W. Chen, and A.L. Ferguson Chem. Sci. 11 9459 (2020)

*This work was supported by MICCoM (Midwest Center for Computational Materials), as part of the Computational Materials Science Program funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Materials Sciences and Engineering Division, and by the National Science Foundation under Grant Nos. CHE-1841805 and ACI-1547580.

–

Presenters

  • Andrew Ferguson

    • University of Chicago
    • Pritzker School of Molecular Engineering, University of Chicago

Authors

  • Andrew Ferguson

    • University of Chicago
    • Pritzker School of Molecular Engineering, University of Chicago

  • Wei Chen

    • University of Illinois

  • Hythem Sidky

    • University of Chicago