AANet: Virtual Screening under Structural Uncertainty via Alignment and Aggregation

2025-06-06

AI Summary

Virtual screening (VS) performance degrades significantly on apo or AlphaFold2-predicted protein structures due to the absence of accurate ligand-binding pocket annotations. Method: We propose the first alignment-aggregation framework explicitly designed for structural uncertainty: a trimodal contrastive learning scheme unifies representations of ligands, holographic pockets, and detected cavities; coupled with a cross-attention adapter that dynamically aggregates candidate binding sites—eliminating reliance on manual pocket annotation while preserving activity signal modeling. Contribution/Results: Evaluated on a newly constructed apo-structure benchmark, our method achieves an enrichment factor EF₁% of 37.19—substantially surpassing the state-of-the-art (11.75). It maintains competitive performance on holo structures, demonstrating strong generalizability and practical utility. This work establishes a novel paradigm for reliable VS in early drug discovery using low-confidence protein structures.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Computer Vision: Other Foundations of Computer Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Bridging structured and unstructured data

Abstract

Virtual screening (VS) is a critical component of modern drug discovery, yet most existing methods--whether physics-based or deep learning-based--are developed around holo protein structures with known ligand-bound pockets. Consequently, their performance degrades significantly on apo or predicted structures such as those from AlphaFold2, which are more representative of real-world early-stage drug discovery, where pocket information is often missing. In this paper, we introduce an alignment-and-aggregation framework to enable accurate virtual screening under structural uncertainty. Our method comprises two core components: (1) a tri-modal contrastive learning module that aligns representations of the ligand, the holo pocket, and cavities detected from structures, thereby enhancing robustness to pocket localization error; and (2) a cross-attention based adapter for dynamically aggregating candidate binding sites, enabling the model to learn from activity data even without precise pocket annotations. We evaluated our method on a newly curated benchmark of apo structures, where it significantly outperforms state-of-the-art methods in blind apo setting, improving the early enrichment factor (EF1%) from 11.75 to 37.19. Notably, it also maintains strong performance on holo structures. These results demonstrate the promise of our approach in advancing first-in-class drug discovery, particularly in scenarios lacking experimentally resolved protein-ligand complexes.

Problem

• Improves virtual screening accuracy for uncertain protein structures
• Aligns ligand, holo pocket, and cavity representations robustly
• Enables learning without precise pocket annotations via dynamic aggregation

Innovation

• Tri-modal contrastive learning aligns ligand and pocket representations
• Cross-attention adapter aggregates candidate binding sites dynamically
• Robust virtual screening under structural uncertainty via alignment-aggregation