Contrastive learning in protein language space predicts interactions between drugs and protein targets
Proceedings of the National Academy of Sciences · 13.06.2023 · 120 (24)
DOI: 10.1073/pnas.2220778120 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the National Academy of Sciences, online erschienen am 08.06.2023. Verfasst von Rohit Singh und 4 weiteren Autor:innen. Beteiligte Einrichtung: Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139. Hauptthemen laut MEDLINE: Proteins, Drug Discovery. Die Arbeit wurde bisher 227-mal zitiert (Crossref) und nennt 83 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: NIGMS NIH HHS, HHS | National Institutes of Health, National Science Foundation.
Abstract (Originalsprache)
Sequence-based prediction of drug–target interactions has the potential to accelerate drug discovery by complementing experimental screens. Such computational prediction needs to be generalizable and scalable while remaining sensitive to subtle variations in the inputs. However, current computational techniques fail to simultaneously meet these goals, often sacrificing performance of one to achieve the others. We develop a deep learning model, ConPLex, successfully leveraging the advances in pretrained protein language models (“PLex”) and employing a protein-anchored contrastive coembedding (“Con”) to outperform state-of-the-art approaches. ConPLex achieves high accuracy, broad adaptivity to unseen data, and specificity against decoy compounds. It makes predictions of binding based on the distance between learned representations, enabling predictions at the scale of massive compound libraries and the human proteome. Experimental testing of 19 kinase-drug interaction predictions validated 12 interactions, including four with subnanomolar affinity, plus a strongly binding EPHB1 inhibitor ( K D = 1.3 nM). Furthermore, ConPLex embeddings are interpretable, which enables us to visualize the drug–target embedding space and use embeddings to characterize the function of human cell-surface proteins. We anticipate that ConPLex will facilitate efficient drug discovery by making highly sensitive in silico drug screening feasible at the genome scale. ConPLex is available open source at https://ConPLex.csail.mit.edu .
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139.
- Ragon Institute of MGH, MIT and Harvard, Cambridge, MA 02139.
- Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139.
- Department of Computer Science, Tufts University, Medford, MA 02155.
- Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA 02139.
ORCID-Profile: Samuel Sledzieski, Bryan Bryson, Lenore Cowen
Förderung
- NIGMS NIH HHS
- HHS | National Institutes of Health
- National Science Foundation
Zitieren
APA
Singh, R., Sledzieski, S., Bryson, B., Cowen, L., & Berger, B. (2023). Contrastive learning in protein language space predicts interactions between drugs and protein targets. Proceedings of the National Academy of Sciences, 120(24). https://doi.org/10.1073/pnas.2220778120
Vancouver
Singh R, Sledzieski S, Bryson B, Cowen L, Berger B. Contrastive learning in protein language space predicts interactions between drugs and protein targets. Proceedings of the National Academy of Sciences. 2023;120(24). doi:10.1073/pnas.2220778120
BibTeX
@article{Singh2023,
author = {Singh, Rohit and Sledzieski, Samuel and Bryson, Bryan and Cowen, Lenore and Berger, Bonnie},
title = {Contrastive learning in protein language space predicts interactions between drugs and protein targets},
journal = {Proceedings of the National Academy of Sciences},
year = {2023},
volume = {120},
number = {24},
doi = {10.1073/pnas.2220778120}
}
EUVIMED