Learning Neural Generative Dynamics for Molecular Conformation Generation

By Minkai Xu, Shitong Luo, Yoshua Bengio, Jian Peng, Jian Tang. 2021

138 citations recorded by the source corpus.

We study how to generate molecule conformations (\textit{i.e.}, 3D structures) from a molecular graph. Traditional methods, such as molecular dynamics, sample conformations via computationally expensive simulations. Recently, machine learning methods have shown great potential by training on a large collection of conformation data. Challenges arise from the limited model capacity for capturing complex distributions of conformations and the difficulty in modeling long-range dependencies between atoms. Inspired by the recent progress in deep generative models, in this paper, we propose a novel probabilistic framework to generate valid and diverse conformations given a molecular graph. We propose a method combining the advantages of both flow-based and energy-based models, enjoying: (1) a high model capacity to estimate the multimodal conformation distribution; (2) explicitly capturing the complex long-range dependencies between atoms in the observation space. Extensive experiments demonstrate the superior performance of the proposed method on several benchmarks, including conformation generation and distance modeling tasks, with a significant improvement over existing generative models for molecular conformation sampling.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology

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 Materials Chemistry

In Molecular Biology

Citations in the corpus

EquiBind: Geometric Deep Learning for Drug Binding Structure Prediction (2022, background): “, 2021) and fast generation of small molecule low-energy conformation ensembles (Ganea et al., 2021b; Luo et al., 2021; Xu et al., 2020; Shi et al., 2021).”
GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation (2022, methodology): “Previous works (Mansimov et al., 2019; Xu et al., 2021b) suggest that this is because RDK IT involves an additional empirical force ﬁeld (FF) (Halgren, 1996) to optimize the structure, and we follow them to also combine G EO D IFF with FF to yield a more fair comparison.”
Torsional Diffusion for Molecular Conformer Generation (2022, background): “A number of machine learning methods for conformer generation has been developed [Xu et al., 2021a,b, Shi et al., 2021, Luo et al., 2021], the most recent and advanced of which are GeoMol [Ganea et al., 2021] and GeoDiff [Xu et al., 2022].”
Learning Gradient Fields for Molecular Conformation Generation (2021, methodology): “Following previous work (Simm & Hernandez-Lobato, 2020; Xu et al., 2021), we report the discrepancy between generated distributions and the ground-truth distributions, by calculating maximum mean discrepancy (MMD) (Gretton et al.”
Predicting Molecular Conformation via Dynamic Graph Score Matching (2021, methodology): “We use the matching score (MAT) to measure the accuracy of generated conformations, and coverage score (COV) to measure the diversity following [34, 43].”
Long Range Graph Benchmark (2022, methodology): “To adapt these tasks to the general graph learning setting, we avoided the prediction of pairwise node distances, which would require specialized methods from the conformer generation literature [21, 52, 63].”
SE(3)-Equivariant Energy-based Models for End-to-End Protein Folding (2021, background): “Xu et al. [47] combine ﬂow-based and energy-based models to predict the inter-atom distance and then recover 3D structures for small molecules.”
Equivariant 3D-conditional diffusion model for molecular linker design (2022, unknown): “To obtain 3D conformations for the molecules generated by DeLinker, we applied a pretrained ConfVAE 45 followed by a force field relaxation procedure MMFF 46 .”
Generating Molecular Conformer Fields (2023, methodology): “In the domain of learning-based approaches, several works have looked at conformer generation problems through the lens of probabilistic modeling, using either normalizing flows (Xu et al., 2021a) or diffusion models (Xu et al., 2022; Jing et al., 2022).”
Crystal Diffusion Variational Autoencoder for Periodic Material Generation (2021, methodology): “Inspired by Xu et al. (2021a); Ganea et al. (2021), we deﬁne two coverage metrics, COV-R (Recall) and COV-P (Precision), to measure the similarity between ensembles of generated materials and ground truth materials in test set.”
GeoMol: Torsional Geometric Generation of Molecular 3D Conformer Ensembles (2021, methodology): “Nevertheless, various facets of the curse of dimensionality have been favorably tackled by ML models in different contexts, and our goal is to build on the recent ML efforts for MCG [Mansimov et al., 2019, Simm and Hernandez-Lobato, 2020, Lemm et al., 2021, Xu et al., 2021].”