GEOM: Energy-annotated molecular conformations for property prediction and molecular generation

By Simon Axelrod, Rafael Gómez‐Bombarelli. 2020

347 citations recorded by the source corpus.

Machine learning outperforms traditional approaches in many molecular design tasks. Although molecules are often thought of as 2D graphs, they in fact consist of an ensemble of inter-converting 3D structures called conformers. Molecular properties arise from the contribution of many conformers, and in the case of a drug binding a target, may be due mainly to a few distinct members. Molecular representations in machine learning are typically based on either one single 3D conformer or on a 2D graph that strips geometrical information. No reference datasets exist that connect these graph and point cloud ensemble representations. Here, we use first-principles simulations to annotate over 400,000 molecules with the ensemble of geometries they span. The Geometrical Embedding Of Molecules (GEOM) dataset contains over 33 million molecular conformers labeled with their relative energies and statistical probabilities at room temperature. This dataset will assist benchmarking and transfer learning in two classes of tasks: inferring 3D properties from 2D molecular graphs, and developing generative models to sample 3D conformations.

Subfields: Computational Theory and Mathematics, Environmental Chemistry, Materials Chemistry

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Environmental Chemistry

In Materials Chemistry

Citations in the corpus

GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation (2022, methodology): “Following prior works (Xu et al., 2021a;b), we also use the recent GEOM-QM9 (Ramakr-ishnan et al., 2014) and GEOM-Drugs (Axelrod & Gomez-Bombarelli, 2020) datasets.”
Learning Gradient Fields for Molecular Conformation Generation (2021, methodology): “We present more visualizations of generated conformations from our ConfGF in Figure 4, including models trained on GEOM-QM9 and GEOM-drugs datasets.”
Uni-Mol: A Universal 3D Molecular Representation Learning Framework (2023, methodology): “We report the mean and median of COV and MAT on GEOM-QM9 and GEOM-Drugs datasets.”
MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, methodology): “GEOM-Drugs [45] also uses advanced sampling and semi-empirical density functional theory to annotate the 3D conformer ensembles.”
Learning Neural Generative Dynamics for Molecular Conformation Generation (2021, methodology): “GEOM-Drugs dataset consists of much larger drug molecules, up to a maximum of 181 atoms (91 heavy atoms).”
Equivariant Diffusion for Molecule Generation in 3D (2022, methodology): “We also measure the energy of generated compounds with the software used to generate the conformations of the GEOM dataset (Bannwarth et al., 2019).”
Pre-training Molecular Graph Representation with 3D Geometry (2021, neutral): “4, 9 [4] Simon Axelrod and Rafael Gomez-Bombarelli.”
Predicting Molecular Conformation via Dynamic Graph Score Matching (2021, methodology): “Following previous works [34, 43], we use the GEOM-QM9 and GEOM-Drugs [1] datasets for this task.”
QMugs, quantum mechanical properties of drug-like molecules (2022, neutral): “…that provide three-dimensional conformational data, for a wide variety of chemical space, at levels of theory higher than classical force fields 32,33 , could boost the performance of machine learning methods in predicting properties from ensembles as well as generative models of conformations.”
Dynamic-Backbone Protein-Ligand Structure Prediction with Multiscale Generative Diffusion Models (2022, neutral): “[56] Simon Axelrod and Rafael Gomez-Bombarelli.”