A geometric deep learning approach to predict binding conformations of bioactive molecules

By Oscar Méndez-Lucio, Mazen Ahmad, E. A. Del Rio-Chanona, Jörg Kurt Wegner. 2021

180 citations recorded by the source corpus.

Subfields: Computational Theory and Mathematics, Molecular Biology, Pharmacology, Graph Learning, Computational Chemistry & Drug Discovery

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. Tommi S. Jaakkola
2. Regina Barzilay
3. Bowen Jing
4. Gabriele Corso
5. Tianfan Fu
6. Jimeng Sun
7. Yuanqi Du
8. Shengchao Liu

In Molecular Biology

1. Tommi S. Jaakkola
2. Regina Barzilay
3. Hannes Stärk
4. Gabriele Corso
5. Bowen Jing
6. Tianfan Fu
7. Zhewei Wei
8. Jiahua Rao

In Pharmacology

1. Jimeng Sun
2. Regina Barzilay
3. Tianfan Fu
4. Tommi S. Jaakkola

In Graph Learning

1. Tommi S. Jaakkola
2. Bowen Jing
3. Regina Barzilay
4. Zhewei Wei
5. Shengchao Liu
6. Gabriele Corso
7. Jimeng Sun
8. Octavian-Eugen Ganea

In Computational Chemistry & Drug Discovery

1. Tommi S. Jaakkola
2. Regina Barzilay
3. Gabriele Corso
4. Bowen Jing
5. Hannes Stärk
6. Zhewei Wei
7. Shengchao Liu
8. Jimeng Sun

Citations in the corpus

• TANKBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction (2022, background): “Many previous works model the interaction between compound and protein by only preserving the interaction region, residues that far away are ignored [26, 44].”
• DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking (2022, methodology): “Recently, machine learning has been applied to parameterize the scoring-function [McNutt et al., 2021; Méndez-Lucio et al., 2021].”
• EquiBind: Geometric Deep Learning for Drug Binding Structure Prediction (2022, methodology): “Closer to our approach, M´endez-Lucio et al. (2021) has shown that optimizing the ligand’s global 3D position and orientation and the torsion angles of rotatable bonds to minimize a GNN based scoring function improves fine-tuning of the ligand into the active site and its predicted bound pose.”
• Artificial intelligence in drug development (2025, unknown): “Current virtual screening models generally focus on specific tasks such as scoring 76 , pose optimization 48 or screening 77 , emphasizing the need to develop universal models capable of handling multiple tasks 78 .”
• MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, methodology): “Representative work : DeepDock [183] designs a model that takes 2D molecular graphs and 2D protein pocket graphs as input and learns the continuous representations.”
• Uni-Mol: A Universal 3D Molecular Representation Learning Framework (2023, methodology): “Also, machine learning-based docking methods, such as ∆ V ina RF 20 [77] and DeepDock [63] have also been developed to predict protein-ligand binding poses and assess protein-ligand binding affinity.”
• Everything is Connected: Graph Neural Networks (2023, background): “, 2020, MaSIF) and Nature Machine Intelligence (Méndez-Lucio et al., 2021).”