MoFlow: An Invertible Flow Model for Generating Molecular Graphs

By Chengxi Zang, Fei Wang. 2020

383 citations recorded by the source corpus.

Generating molecular graphs with desired chemical properties driven by deep graph generative models provides a very promising way to accelerate drug discovery process. Such graph generative models usually consist of two steps: learning latent representations and generation of molecular graphs. However, to generate novel and chemically-valid molecular graphs from latent representations is very challenging because of the chemical constraints and combinatorial complexity of molecular graphs. In this paper, we propose MoFlow, a flow-based graph generative model to learn invertible mappings between molecular graphs and their latent representations. To generate molecular graphs, our MoFlow first generates bonds (edges) through a Glow based model, then generates atoms (nodes) given bonds by a novel graph conditional flow, and finally assembles them into a chemically valid molecular graph with a posthoc validity correction. Our MoFlow has merits including exact and tractable likelihood training, efficient one-pass embedding and generation, chemical validity guarantees, 100% reconstruction of training data, and good generalization ability. We validate our model by four tasks: molecular graph generation and reconstruction, visualization of the continuous latent space, property optimization, and constrained property optimization. Our MoFlow achieves state-of-the-art performance, which implies its potential efficiency and effectiveness to explore large chemical space for drug discovery.

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

1. Tianfan Fu
2. Sung Ju Hwang
3. Jaehyeong Jo
4. Seul Lee
5. Heng Ji
6. Jiawei Han
7. Shengchao Liu
8. Dejing Dou

In Materials Chemistry

1. Heng Ji
2. Meng Jiang
3. Shengchao Liu
4. Jimeng Sun
5. Anima Anandkumar
6. Weili Nie
7. Jaehyeong Jo
8. Seul Lee

In Molecular Biology

1. Sung Ju Hwang
2. Jaehyeong Jo
3. Seul Lee
4. Tianfan Fu
5. Heng Ji
6. Meng Jiang
7. Jiawei Han
8. Shengchao Liu

Citations in the corpus

• State-specific protein–ligand complex structure prediction with a multiscale deep generative model (2022, positive): “A closed-loop combination of NeuralPLexer and recent advances in differentiable generative models for protein sequences 27,28 , molecular graphs 57,58 and structure-based bioactivity models holds the promise of accelerating the process of sequence and compound prioritization over conventional…”
• Dynamic-Backbone Protein-Ligand Structure Prediction with Multiscale Generative Diffusion Models (2022, background): “…NeuralPLexer only requires molecular graphs as ligand inputs, therefore can enable end-to-end gradient-based design for functional small-molecules and ligand-binding proteins when coupled to recently-proposed differentiable protein sequence [24–26] and molecular graph generators [27, 28].”
• Large Language Models on Graphs: A Comprehensive Survey (2023, methodology): “However, the aforementioned methods cannot leverage task labels.”
• MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, background): “MoFlow [130] designs a flow-based model to generate molecular graphs in a one-shot manner, which builds the node feature and edge feature matrices as a whole.”
• Differentiable Scaffolding Tree for Molecular Optimization (2021, neutral): “Typical algorithms include variational autoencoder (VAE), generative adversarial network (GAN), energy-based models, flow-based model [14, 21, 9, 37, 23, 17, 34, 38, 22, 49, 27, 5, 32, 1].”
• Structure-aware Interactive Graph Neural Networks for the Prediction of Protein-Ligand Binding Affinity (2021, background): “By contrast, many researchers have greatly developed GNN models in other fields of drug discovery [37, 45], such as predicting molecular property [17, 26, 44] and chemical reaction [7].”
• Score-based Generative Modeling of Graphs via the System of Stochastic Differential Equations (2022, methodology): “…or edge resampling, which is different from the metric used in Shi et al. (2020) and Luo et al. (2021), since we allow atoms to have formal charge when checking their valency following Zang & Wang (2020), which is more reasonable due to the existence of formal charge in the training molecules.”