Macao Polytechnic University Research Team Develops Intelligent Drug Design Method, Advancing Molecular Generation Technology in AI-Driven Drug Discovery
Macao Polytechnic University Research Team Develops Intelligent Drug Design Method, Advancing Molecular Generation Technology in AI-Driven Drug Discovery
Research findings from Macao Polytechnic University were published in the prestigious international journal Advanced Science (Image Source: Advanced Science)
Professors Liu Huanxiang and Yao Xiaojun from the Center for Artificial Intelligence Driven Drug Discovery of Macao Polytechnic University led a research team in publishing an academic paper titled "DiffMC-Gen: A Dual Denoising Diffusion Model for Multi-Conditional Molecular Generation" in the internationally renowned journal Advanced Science. The first author of the paper is Yang Yuwei, a second-year Ph.D. student in the PhD Program in Artificial Intelligence Driven Drug Discovery at Macao Polytechnic University.
This research addresses the long challenge of designing small-molecule drugs with diverse physicochemical properties in a precise and efficient manner. The team developed DiffMC-Gen, an innovative molecular generative model, integrates discrete and continuous diffusion models while embedding both 2D and 3D molecular features. DiffMC-Gen generates novel molecules that meet essential drug design criteria by optimizing pharmacophore matching coefficient, drug-likeness, synthetic accessibility, and toxicity risk. This research contributes to the balance of drug-likeness and synthetic feasibility while enhancing target affinity and structural novelty in drug discovery. It provides a new approach for the technical evolution of AI-driven drug discovery.
DiffMC-Gen: A Dual Denoising Diffusion Model for Multi-Conditional Molecular Generation.
Advanced Science is a globally recognized, interdisciplinary international science journal published by Wiley. It is dedicated to publishing best-in-class fundamental and applied research in materials science, physics, chemistry, medicine, life sciences, and engineering. The journal consistently ranks within the top 10% across multiple disciplines. It is classified as a CAS Q1 and JCR Q1 journal, with a five-year impact factor of 16.3 and a 2023 impact factor of 14.3.
This study is supported by the Science and Technology Development Fund, Macau, SAR (No. 0043/2023/AFJ) and Macao Polytechnic University (No. RP/ FCA-02 / 2023).