🚀 Redefining Drug-Likeness with AI: ICML & ISMB Presentations Coming Up! 💊
“Can this molecule become a drug?” 🤔
That’s the ultimate question in drug discovery, but for over 20 years, we’ve been relying on Lipinski’s Rule of 5 and other old heuristics.
✨ This July, we’re bringing AI to redefine drug-likeness at two of the world’s top AI & bioinformatics conferences:
📍 ICML 2025 (Vancouver, July 15)
“BounDr.E: Predicting Drug-likeness via Biomedical Knowledge Alignment and EM-like One-Class Boundary Optimization.” We propose an EM-like one-class boundary optimization approach, aligning biomedical knowledge to predict drug-likeness without heuristic rules.
📍 ISMB/ECCB 2025 (Liverpool, July 24)
“ADME-Drug-Likeness: Enriching Molecular Foundation Models via Pharmacokinetics-Guided Multi-Task Learning for Drug-likeness Prediction.” We integrate ADME (absorption, distribution, metabolism, excretion) properties into molecular foundation models to enhance the interpretability and predictive power of drug-likeness.
Our mission at AIGENDRUG and SNU (Prof. Sun Kim’s lab) is to bring data-driven, quantitative definitions of “drug-likeness” into the early stages of drug discovery, moving beyond rules, and enabling smarter drug design. 💡
Looking forward to sharing these ideas with the global AI and bioinformatics community this summer! 🌍🧬