Uni-Mol: A Universal 3D Molecular Representation Learning Framework

Abstract

Molecular representation learning (MRL) has gained tremendous attention due to its critical role in learning from limited supervised data for applications like drug design. In most MRL methods, molecules are treated as 1D sequential tokens or 2D topology graphs, limiting their ability to incorporate 3D information for downstream tasks and, in particular, making it almost impossible for 3D geometry prediction or generation. Herein, we propose Uni-Mol, a universal MRL framework that significantly enlarges the representation ability and application scope of MRL schemes. Uni-Mol is composed of two models with the same SE(3)-equivariant transformer architecture: a molecular pretraining model trained by 209M molecular conformations; a pocket pretraining model trained by 3M candidate protein pocket data. The two models are used independently for separate tasks, and are combined when used in protein-ligand binding tasks. By properly incorporating 3D information, Uni-Mol outperforms SOTA in 14/15 molecular property prediction tasks. Moreover, Uni-Mol achieves superior performance in 3D spatial tasks, including protein-ligand binding pose prediction, molecular conformation generation, etc. Finally, we show that Uni-Mol can be successfully applied to the tasks with few-shot data like pocket druggability prediction. The code, model, and data are made publicly available at \url{https://github.com/dptech-corp/Uni-Mol}.

OLB-AC: toward optimizing ligand bioactivities through deep graph learning and activity cliffs

“…It then minimizes the maximum loss of prediction divergence on bioactivity to ensure the local smoothness of model outputs ( Yin et al 2022a ). Uni-Mol is a universal 3D molecular representation learning framework that significantly enlarges the representation ability and application scope of molecular representation learning schemes ( Zhou et al 2023 ).…”

“… Li et al (2022a ) proposed KPGT, a new knowledge-guided pre-training Graph Transformer model, a novel self-supervised learning framework for molecular graph representation learning. Zhou et al (2023) deployed a universal molecular representation learning framework that enlarges the representation ability and application scope of molecular representation learning schemes including ligand bioactivity prediction.…”

ArtiDock: fast and accurate machine learning approach to protein-ligand docking based on multimodal data augmentation

“…For the point cloud generation we combined and improved the ligand pose generation approaches from TankBind 3,26,27 and Uni-Mol 28 . Given the model predicted , we 𝐷𝑀 𝑝𝑐 randomly initialized the 3D point cloud , where c -number of ligand atoms, and 𝐶' ∈ ℝ 𝑐×3 applied back-propagation to optimize the by Adam optimizer.…”

Stereochemically-aware bioactivity descriptors for uncharacterized chemical compounds

“…Signaturizers3D, Fig 2e ). We first generated 3D conformations for all CC molecules, coupled them with their type III signatures, and used them to fine-tune the pre-trained Uni-Mol model 17 (see Supplementary Information ). We then evaluated the capability of Signaturizers3D to distinguish stereoisomers by generating B4 signatures for the 32,705 compounds identified in the CC B4 space and calculating distances between stereoisomer pairs.…”