Peer-Reviewed Publication
J Comput Aided Mol Des2026;40(1):58.February 5, 2026Journal Article
Comparing massively-multitask regression algorithms for drug discovery.
Eric J Martin1, Xiang-Wei Zhu2, Patrick Riley3, Steven Kearnes4, Ekaterina A Sosnina5, Li Tian6,7, Chi-Ming Che6,7, Zijian Wang6,7, Ying Wei8, Thomas M Whitehead9, Gareth J Conduit9, Matthew D Segall10
1Novartis Biomedical Research, Emeryville, CA, 94608, USA. eric.martin@novartis.com.
2Novartis Biomedical Research, Emeryville, CA, 94608, USA.
3DaVinci, Cambridge, MA, USA.
4Genesis Therapeutics, Burlingame, CA, 94010, USA.
5The Medical University of Innsbruck, Innsbruck, Austria.
6Laboratory for Synthetic Chemistry and Chemical Biology Limited, Hong Kong Science Park, Shatin, Hong Kong, China.
7State Key Laboratory of Synthetic Chemistry and Department of Chemistry, The University of Hong Kong, Pokfulam, Hong Kong, China.
8College of Computer Science and Technology, Zhejiang University, Hangzhou, China.
9Intellegens Limited, Chesterton Mill, Cambridge, CB4 3NP, UK.
10Optibrium Limited, Cambridge Innovation Park, Denny End Rd, Cambridge, CB25 9PB, UK.
Abstract
Massively-multitask regression models (MMRMs) have revolutionized activity prediction for drug discovery. MMRMs trained on millions of compounds and many thousands of assays can predict bioactivity with accuracy comparable to 4-concentration IC50 experiments. This report compares six MMRMs: pQSAR, Alchemite, MT-DNN, MetaNN, Macau and IMC. Models were trained by experts in each method, on identical…
Create a free account to keep reading
Free members get 10 full research views every month across publications, clinical trials, FDA clearances, adverse events, and NIH grants. No credit card required.
Want unlimited research access? See Pro plans