Chemically Transferable Generative Backmapping of Coarse-Grained Proteins | Soojung Yang
Try datamol.io - the open source toolkit that simplifies molecular processing and featurization workflows for machine learning scientists working in drug discovery: https://datamol.io/
Never miss another LoGG Talk, add the schedule to your calendar: https://m2d2.io/talks/m2d2/about/
Join the Learning on Graphs and Geometry Reading Group on Slack: https://join.slack.com/t/logag/shared...
Abstract: Coarse-graining (CG) accelerates molecular simulations of protein dynamics by simulating sets of atoms as singular beads. Backmapping is the opposite operation of bringing lost atomistic details back from the CG representation. While machine learning (ML) has produced accurate and efficient CG simulations of proteins, fast and reliable backmapping remains a challenge. Rule-based methods produce poor all-atom geometries, needing computationally costly refinement through additional simulations. Recently proposed ML approaches outperform traditional baselines but are not transferable between proteins and sometimes generate unphysical atom placements with steric clashes and implausible torsion angles. This work addresses both issues to build a fast, transferable, and reliable generative backmapping tool for CG protein representations. We achieve generalization and reliability through a combined set of innovations: representation based on internal coordinates; an equivariant encoder/prior; a custom loss function that helps ensure local structure, global structure, and physical constraints; and expert curation of high-quality out-of-equilibrium protein data for training. Our results pave the way for out-of-the-box backmapping of coarse-grained simulations for arbitrary proteins.
Speakers: Soojung Yang - https://sites.google.com/view/soojungy/
Twitter Hannes: / hannesstaerk
Twitter Dominique: / dom_beaini
Twitter datamol.io: / datamol_io
~
Chapters
00:00 - Intro
01:26 - Glossary
03:25 - Protein Conformation and Energy Landscape
06:30 - Seminal Prior Work
07:19 - Conditional VAE Framework of Generative Backmapping
09:01 - Can Backmapping be Made Transferable?
11:40 - Proof of Concept Study
13:28 - Data: Protein Ensemble Database
15:46 - Encoder and Prior: Equivariant Message Passing
24:23 - Decoder: Backbone Reconstruction
27:21 - Encoder: Sidechain Reconstruction
31:37 - Learning Objectives
35:32 - Ablation Study 1: Transferability
38:31 - Ablation Study 2: Equivariance of the Encoder
39:50 - Ablation Study 3: Decoder Degrees of Freedom
46:22 - Diversity Metrics
57:14 - Q+A