[Other] [Other]EQUIBIND: Geometric Deep Learning for Drug Binding Structure Prediction

water56965057 Post time 1 hour(s) ago | Show all posts |Read mode
This post will be closed automatically in 2026-07-26 13:28
Reward10points

ICLR 2022 Proceedings PaperㄛarXiv:2202.05146
ByㄩStärk, H (Staerk, Hannes) [1] ; Ganea, OE (Ganea, Octavian-Eugen) [1] ; Pattanaik, L (Pattanaik, Lagnajit) [1] ; Barzilay, R (Barzilay, Regina) [1] ; Jaakkola, T (Jaakkola, Tommi) [1]
Edited byㄩChaudhuri, K (Chaudhuri, K) ; Jegelka, S (Jegelka, S) ; Song, L (Song, L) ; Szepesvari, C (Szepesvari, C) ; Niu, G (Niu, G) ; Sabato, S (Sabato, S)
SourceㄩINTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 162
Book SeriesㄩProceedings of Machine Learning Research
Publishedㄩ2022
Indexedㄩ2023-03-15
AbstractㄩPredicting how a drug-like molecule binds to a specific protein target is a core problem in drug discovery. An extremely fast computational binding method would enable key applications such as fast virtual screening or drug engineering. Existing methods are computationally expensive as they rely on heavy candidate sampling coupled with scoring, ranking, and fine-tuning steps. We challenge this paradigm with EQUIBIND, an SE(3)-equivariant geometric deep learning model performing direct-shot prediction of both i) the receptor binding location (blind docking) and ii) the ligand's bound pose and orientation. EquiBind achieves significant speed-ups and better quality compared to traditional and recent baselines. Further, we show extra improvements when coupling it with existing fine-tuning techniques at the cost of increased running time. Finally, we propose a novel and fast fine-tuning model that adjusts torsion angles of a ligand's rotatable bonds based on closed-form global minima of the von Mises angular distance to a given input atomic point cloud, avoiding previous expensive differential evolution strategies for energy minimization.
Reply

Use magic Donate Report

All Reply0 Show all posts

Reply

You have to log in before you can reply Login | Register

Points Rules

Return to the list