Geometric Deep Learning for Molecular Crystal Structure Prediction
Authors:
Preprint
Abstract
We develop and test new machine learning strategies for accelerating molecular crystal structure ranking and crystal property prediction using tools from geometric deep learning on molecular graphs. Leveraging developments in graph-based learning and the availability of large molecular crystal datasets, we train models for density prediction and stability ranking which are accurate, fast to evaluate, and applicable to molecules of widely varying size and composition. Our density prediction model, MolXtalNet-D, achieves state of the art performance, with lower than 2% mean absolute error on a large and diverse test dataset. Our crystal ranking tool, MolXtalNet-S, correctly discriminates experimental samples from synthetically generated fakes and is further validated through analysis of the submissions to the Cambridge Structural Database Blind Tests 5 and 6. Our new tools are computationally cheap and flexible enough to be deployed within an existing crystal structure prediction pipeline both to reduce the search space and score/filter crystal candidates.Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Symmetric Molecular DynamicsJournal of Chemical Theory and Computation, 2022
- Machine Learned Model for Solid Form Volume Estimation Based on Packing-Accessible Surface and Molecular Topological FragmentsThe Journal of Physical Chemistry A, 2020
- PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments, and Partial ChargesJournal of Chemical Theory and Computation, 2019
- Graph Networks as a Universal Machine Learning Framework for Molecules and CrystalsChemistry of Materials, 2019
- GAtor: A First-Principles Genetic Algorithm for Molecular Crystal Structure PredictionJournal of Chemical Theory and Computation, 2018
- DDT Polymorphism and the Lethality of Crystal FormsAngewandte Chemie, 2017
- Development and testing of a general amber force fieldJournal of Computational Chemistry, 2004
- Upack program package for crystal structure prediction: Force fields and crystal structure generation for small carbohydrate moleculesJournal of Computational Chemistry, 1999
- Crystal Structure Prediction based on Statistical PotentialsJournal of Molecular Modeling, 1998
- IsoStar: A library of information about nonbonded interactionsJournal of Computer-Aided Molecular Design, 1997