Peer-Reviewed Publication
Science2025;390(6776):eadi8577.November 27, 2025Journal Article
Active learning framework leveraging transcriptomics identifies modulators of disease phenotypes.
Benjamin DeMeo1, Charlotte Nesbitt1, Samuel A Miller1, Daniel B Burkhardt1, Inna Lipchina1, Doris Fu1, Peter Holderrieth1, David Kim1, Sergey Kolchenko1, Artur Szalata2,3, Ishan Gupta1, Christine Kerr1, Thomas Pfefer1, Raziel Rojas-Rodriguez1, Sunil Kuppassani1, Laurens Kruidenier1, Parul B Doshi1, Mahdi Zamanighomi1, James J Collins4,5,6,7, Alex K Shalek4,6,8,9, Fabian J Theis2,3,10, Mauricio Cortes1
1Cellarity Inc, Somerville, MA, USA.
2Computational Health Center, Institute of Computational Biology, Helmholtz-Munich, Neuherberg, Germany.
3TUM School of Computation, Information and Technology, Technical University of Munich, Garching, Germany.
4Institute for Medical Engineering and Science, Massachusetts Institute of Technology (MIT), Cambridge, MA, USA.
5Department of Biological Engineering, MIT Cambridge, MA, USA.
6Broad Institute of MIT and Harvard, Cambridge, MA, USA.
7Wyss Institute, Harvard University, Boston, MA, USA.
8Department of Chemistry and Koch Institute for Integrative Cancer Research, MIT Cambridge, MA, USA.
9Ragon Institute of MGH, MIT, and Harvard, Cambridge, MA, USA.
10Helmholtz AI, Helmholtz-Munich, Neuherberg, Germany.
Abstract
Phenotypic drug screening remains constrained by the vastness of chemical space and the technical challenges of scaling experimental workflows. To overcome these barriers, computational methods have been developed to prioritize compounds, but they rely on either single-task models lacking generalizability or heuristic-based genomic proxies that resist optimization. We designed an active deep learn…
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