Expert-Inspired Multi-Agent Coordination for Multi-Objective Molecular Optimization
2026-03-14 Proceedings of the AAAI Conference on Artificial Intelligence (AAAI-26)

Daojian Zeng, Tianle Li, Jiahao Yang, Jiacai Yi, Xieping Gao, Lincheng Jiang, Tengfei Ma, Xiangxiang Zeng

MAMO is a multi-agent molecular design framework that simulates expert collaboration. Specialized agents optimize individual objectives while a central scheduler reallocates tasks from evaluation feedback, balancing conflicting properties.

Link
Decoding the limits of deep learning in molecular docking for drug discovery
2025-10-01 Chemical Science

Yue Li, Jiacai Yi, Hui Li, Kun Li, Fenghua Kang, Youchao Deng, Chengkun Wu, Xiangzheng Fu, Dejun Jiang, Dongsheng Cao

Structure-based molecular docking, a cornerstone of computational drug design, is undergoing a paradigm shift fueled by deep learning (DL) innovations. However, the rapid proliferation of DL-driven docking methods…

Link
DeepMetab: a comprehensive and mechanistically informed graph learning framework for end-to-end drug metabolism prediction
2025-09-05 Chemical Science

Yiling Zhou, Dejun Jiang, Xiao Wei, Jiacai Yi, Yikun Wang, Youchao Deng, Dongsheng Cao

DeepMetab is the first comprehensive, mechanistically informed deep graph learning framework for end-to-end prediction of CYP450-mediated drug metabolism, unifying substrate profiling, site-of-metabolism localization, and metabolite generation in a single multi-task architecture…

Link
ADMETlab 3.0: an updated comprehensive online ADMET prediction platform enhanced with broader coverage, improved performance, API functionality and decision support
2024-07-01 Nucleic Acids Research

Li Fu, Shaohua Shi, Jiacai Yi, Ningning Wang, Yuanhang He, Zhenxing Wu, Jinfu Peng, Youchao Deng, Wenxuan Wang, Chengkun Wu, Aiping Lyu, Xiangxiang Zeng, Wentao Zhao, Tingjun Hou, Dongsheng Cao

ADMETlab 3.0 is the second updated version of the web server that provides a comprehensive and efficient platform for evaluating ADMET-related parameters as well as physicochemical properties and medicinal chemistry…

Link
ChemFH: an integrated tool for screening frequent false positives in chemical biology and drug discovery
2024-07-01 Nucleic Acids Research

Shaohua Shi, Li Fu, Jiacai Yi, Ziyi Yang, Xiaochen Zhang, Youchao Deng, Wenxuan Wang, Chengkun Wu, Wentao Zhao, Tingjun Hou, Xiangxiang Zeng, Aiping Lyu, Dongsheng Cao

High-throughput screening rapidly tests extensive arrays of chemical compounds to identify hit compounds for specific biological targets in drug discovery. However, false-positive results disrupt hit-to-lead progression…

Link
OptADMET: a web-based tool for substructure modifications to improve ADMET properties of lead compounds
2024-04-01 Nature Protocols

Jiacai Yi, Shaohua Shi, Li Fu, Ziyi Yang, Pengfei Nie, Aiping Lu, Chengkun Wu, Yafeng Deng, Changyu Hsieh, Xiangxiang Zeng, Tingjun Hou, Dongsheng Cao

Lead optimization is a crucial step in drug discovery, aiming to design potential drug candidates from biologically active hits. During lead optimization, active hits undergo modifications to achieve a balance between…

Link
ADMETlab 2.0: an integrated online platform for accurate and comprehensive predictions of ADMET properties
2021-07-01 Nucleic Acids Research

Guoli Xiong, Zhenxing Wu, Jiacai Yi, Li Fu, Zhijiang Yang, Changyu Hsieh, Mingzhu Yin, Xiangxiang Zeng, Chengkun Wu, Aiping Lu, Xiang Chen, Tingjun Hou, Dongsheng Cao

Because undesirable pharmacokinetics and toxicity of candidate compounds are the main reasons for the failure of drug development, it has been widely recognized that ADMET should be evaluated as early as possible. Here,…

Link
ADMETlab 2.0

An integrated online platform for accurate and comprehensive ADMET property prediction. - Established an earlier large-scale online ADMET prediction platform. - Supported broad property prediction for medicinal…

Link
ADMETlab 3.0

A comprehensive online ADMET prediction platform with API access and decision support for drug discovery. - Covers 119 endpoints and more than 400,000 entries according to the CV. - Built on a multitask DMPNN framework…

Link