Paper List

期刊: Bioinformatics
发布日期: 2026-03-16
BioinformaticsDrug Discovery

Empowering Chemical Structures with Biological Insights for Scalable Phenotypic Virtual Screening

Hunan University | Jiangnan University | University of Tsukuba | Hong Kong Baptist University | Xiamen University

Xiaoqing Lian, Pengsen Ma, Tengfeng Ma, Zhonghao Ren, Xibao Cai, Zhixiang Cheng, Bosheng Song, He Wang, Xiang Pan, Yangyang Chen, Sisi Yuan, Chen Lin
Figure
Figure
Figure
Figure
Figure

30秒速读

IN SHORT: This paper addresses the core challenge of bridging the gap between scalable chemical structure screening and biologically informative but resource-intensive phenotypic profiling in drug discovery.

核心创新

  • Methodology Introduces DECODE framework that uses geometric disentanglement to separate measurement-invariant biological signals from modality-specific experimental noise
  • Methodology Implements contrastive learning with orthogonal constraints to align heterogeneous biological modalities into a unified biological consensus space
  • Methodology Develops three adaptive inference protocols including zero-shot retrieval, dynamic adaptation, and generative integration for virtual screening

主要结论

  • DECODE achieves over 20% relative improvement in zero-shot MOA prediction compared to chemical baselines, demonstrating effective biological signal extraction from structures alone
  • The framework yields a 6-fold increase in hit rates for novel anti-cancer agents during external validation (AUC: 0.737 vs 0.694 for chemical baseline)
  • DECODE's disentanglement mechanism improves F1-score by 15.8% over expert MLP baselines on sparsely labeled CDRP dataset, showing robustness against experimental noise
研究空白： Current drug discovery faces a fundamental trade-off: chemical structure screening is scalable but lacks biological context, while phenotypic profiling provides biological insights but is resource-intensive. Existing methods struggle to extract robust biological signals from noisy data and encode them into representations that don't require biological data at inference.

摘要: Motivation: The scalable identification of bioactive compounds is essential for contemporary drug discovery. This process faces a key trade-off: structural screening offers scalability but lacks biological context, whereas high-content phenotypic profiling provides deep biological insights but is resource-intensive. The primary challenge is to extract robust biological signals from noisy data and encode them into representations that do not require biological data at inference. Results: This study presents DECODE (DEcomposing Cellular Observations of Drug Effects), a framework that bridges this gap by empowering chemical representations with intrinsic biological semantics to enable structure-based in silico biological profiling. DECODE leverages limited paired transcriptomic and morphological data as supervisory signals during training, enabling the extraction of a measurement-invariant biological fingerprint from chemical structures and explicit filtering of experimental noise. Our evaluations demonstrate that DECODE retrieves functionally similar drugs in zero-shot settings with over 20% relative improvement over chemical baselines in mechanism-of-action (MOA) prediction. Furthermore, the framework achieves a 6-fold increase in hit rates for novel anti-cancer agents during external validation. Availability and implementation: The codes and datasets of DECODE are available at https://github.com/lian-xiao/DECODE.

代码