Pearl: A Foundation Model for Placing Every Atom in the Right Location

2025-10-28

AI Summary

Predicting the 3D structures of protein–ligand complexes remains a critical bottleneck in computational drug discovery. To address this, we introduce the first large-scale SO(3)-equivariant diffusion foundation model tailored for joint protein–ligand folding, designed to enhance both prediction accuracy and physical plausibility. Methodologically: (1) we construct a high-quality synthetic dataset to mitigate scarcity of experimentally resolved complex structures; (2) we develop an SO(3)-equivariant diffusion architecture that rigorously enforces rotational symmetry; and (3) we propose a multi-chain templated controllable inference mechanism enabling co-modeling of proteins and non-polymeric ligands. Experiments demonstrate substantial improvements over AlphaFold 3: +14.5% and +14.2% in the fraction of predictions achieving RMSD < 2 Å on Runs N’Poses and PoseBusters benchmarks, respectively, and a 3.6× gain under the stricter RMSD < 1 Å threshold. These advances significantly advance high-accuracy, physically interpretable drug discovery.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Diffusion Models for Vision
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Accurately predicting the three-dimensional structures of protein-ligand complexes remains a fundamental challenge in computational drug discovery that limits the pace and success of therapeutic design. Deep learning methods have recently shown strong potential as structural prediction tools, achieving promising accuracy across diverse biomolecular systems. However, their performance and utility are constrained by scarce experimental data, inefficient architectures, physically invalid poses, and the limited ability to exploit auxiliary information available at inference. To address these issues, we introduce Pearl (Placing Every Atom in the Right Location), a foundation model for protein-ligand cofolding at scale. Pearl addresses these challenges with three key innovations: (1) training recipes that include large-scale synthetic data to overcome data scarcity; (2) architectures that incorporate an SO(3)-equivariant diffusion module to inherently respect 3D rotational symmetries, improving generalization and sample efficiency, and (3) controllable inference, including a generalized multi-chain templating system supporting both protein and non-polymeric components as well as dual unconditional/conditional modes. Pearl establishes a new state-of-the-art performance in protein-ligand cofolding. On the key metric of generating accurate (RMSD<2 {A}) and physically valid poses, Pearl surpasses AlphaFold 3 and other open source baselines on the public Runs N'Poses and PoseBusters benchmarks, delivering 14.5% and 14.2% improvements, respectively, over the next best model. In the pocket-conditional cofolding regime, Pearl delivers $3.6 imes$ improvement on a proprietary set of challenging, real-world drug targets at the more rigorous RMSD<1 {A} threshold. Finally, we demonstrate that model performance correlates directly with synthetic dataset size used in training.

Problem

• Predicting 3D protein-ligand complex structures accurately
• Overcoming data scarcity and inefficient architectures in prediction
• Generating physically valid poses for computational drug discovery

Innovation

• Large-scale synthetic data training to overcome scarcity
• SO(3)-equivariant diffusion module for 3D symmetries
• Controllable inference with multi-chain templating system