EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Assessing the Generalizability of Machine Learning and Physics-Based Methods with DNA-Encoded Libraries

Marissa Dolorfino, Daniel Santos Perez, Yao Fu, Shu-Hang Lin iD, Sean McCarty, Matthew J O'Meara, Terra Sztain iD

Journal of Chemical Information and Modeling · 2026

DOI: 10.1021/acs.jcim.6c01611 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Chemical Information and Modeling, 2026. Verfasst von Marissa Dolorfino und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Medicinal Chemistry, University of Michigan, Ann Arbor, Michigan48104, United States. Hauptthemen laut MEDLINE: Machine Learning, DNA, Drug Discovery, Small Molecule Libraries. Die Arbeit nennt 75 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: NIGMS NIH HHS, U.S. Department of Energy, National Institute of General Medical Sciences.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Predicting protein–ligand binding is a central challenge in computational drug discovery, and while machine learning (ML) and cofolding methods have advanced rapidly, their ability to generalize beyond training or parametrization regimes remains insufficiently understood. DNA-encoded libraries (DELs) enable ultralarge screening of billions of molecules simultaneously, providing a useful testbed for evaluating these approaches at scale. A recent NeurIPS competition revealed that even top-performing ML models trained on DEL data failed at generalizing to out-of-distribution (OOD) chemical space. We investigated whether integrating structural modeling could bridge this generalization gap. We systematically assessed state-of-the-art ML, docking, and cofolding methods, including Schrödinger Glide, Rosetta GALigandDock, and Boltz-2 with three biologically diverse protein targets screened against libraries containing multiple DEL synthesis formats. While ML excels in-distribution, OOD hit discrimination is dependent on both the target and ligand context, with no single method consistently dominating. These findings demonstrate that benchmark performance alone is insufficient to predict OOD performance, highlighting the need for system-dependent evaluation of binding prediction methods. We provide an open-source package for assessing protein–ligand prediction methods and analyzing high-throughput screening data: DEL-iver.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Machine LearningDNADrug DiscoverySmall Molecule LibrariesProteinsLigandsMolecular Docking Simulation

Substanzen

DNAProteinsLigandsSmall Molecule Libraries

Autor:innen und Institutionen

  • Department of Medicinal Chemistry, University of Michigan, Ann Arbor, Michigan48104, United States.
  • Gilbert S. Omenn Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, Michigan48109, United States.
  • Program in Biophysics, University of Michigan, Ann Arbor, Michigan48109, United States.

ORCID-Profile: Shu-Hang Lin, Terra Sztain

Förderung

  • NIGMS NIH HHS
  • U.S. Department of Energy
  • National Institute of General Medical Sciences

Zitieren

APA

Dolorfino, M., Santos Perez, D., Fu, Y., Lin, S. H., McCarty, S., O'Meara, M. J., & Sztain, T. (2026). Assessing the Generalizability of Machine Learning and Physics-Based Methods with DNA-Encoded Libraries. Journal of Chemical Information and Modeling. https://doi.org/10.1021/acs.jcim.6c01611

Vancouver

Dolorfino M, Santos Perez D, Fu Y, Lin SH, McCarty S, O'Meara MJ, et al. Assessing the Generalizability of Machine Learning and Physics-Based Methods with DNA-Encoded Libraries. Journal of Chemical Information and Modeling. 2026. doi:10.1021/acs.jcim.6c01611

BibTeX
@article{Dolorfino2026,
  author = {Dolorfino, Marissa and Santos Perez, Daniel and Fu, Yao and Lin, Shu-Hang and McCarty, Sean and O'Meara, Matthew J and Sztain, Terra},
  title = {Assessing the Generalizability of Machine Learning and Physics-Based Methods with DNA-Encoded Libraries},
  journal = {Journal of Chemical Information and Modeling},
  year = {2026},
  doi = {10.1021/acs.jcim.6c01611}
}
Kostenlos starten →