EUVIMEDEUVIMED entdecken
Crossref

Adapting Deep Learning QSPR Models to Specific Drug Discovery Projects

Andrin Fluetsch, Elena Di Lascio, Grégori Gerebtzoff, Raquel Rodríguez-Pérez iD

Molecular Pharmaceutics · 01.04.2024 · 21 (4) · 1817-1826

DOI: 10.1021/acs.molpharmaceut.3c01124 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Molecular Pharmaceutics, online erschienen am 19.02.2024. Verfasst von Andrin Fluetsch und 3 weiteren Autor:innen. Beteiligte Einrichtung: Novartis Biomedical Research, Novartis Campus, Basel 4002, Switzerland. Hauptthemen laut MEDLINE: Deep Learning. Die Arbeit wurde bisher 18-mal zitiert (Crossref) und nennt 45 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Medicinal chemistry and drug design efforts can be assisted by machine learning (ML) models that relate the molecular structure to compound properties. Such quantitative structure-property relationship models are generally trained on large data sets that include diverse chemical series (global models). In the pharmaceutical industry, these ML global models are available across discovery projects as an "out-of-the-box" solution to assist in drug design, synthesis prioritization, and experiment selection. However, drug discovery projects typically focus on confined parts of the chemical space (e.g., chemical series), where global models might not be applicable. Local ML models are sometimes generated to focus on specific projects or series. Herein, ML-based global models, local models, and hybrid global-local strategies were benchmarked. Analyses were done for more than 300 drug discovery projects at Novartis and ten absorption, distribution, metabolism, and excretion (ADME) assays. In this work, hybrid global-local strategies based on transfer learning approaches were proposed to leverage both historical ADME data (global) and project-specific data (local) to adapt model predictions. Fine-tuning a pretrained global ML model (used for weights' initialization, WI) was the top-performing method. Average improvements of mean absolute errors across all assays were 16% and 27% compared with global and local models, respectively. Interestingly, when the effect of training set size was analyzed, WI fine-tuning was found to be successful even in low-data scenarios (e.g., ∼10 molecules per project). Taken together, this work highlights the potential of domain adaptation in the field of molecular property predictions to refine existing pretrained models on a new compound data distribution.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningDrug DiscoveryDrug DesignMachine LearningQuantitative Structure-Activity Relationship

Stichwörter der Autor:innen

ADMEdeep learningdomain adaptationdrug designlocal modelsmachine learningquantitative structure−property relationships (QSPR)transfer learning

Autor:innen und Institutionen

  • Novartis Biomedical Research, Novartis Campus, Basel 4002, Switzerland.

ORCID-Profile: Raquel Rodríguez-Pérez

Zitieren

APA

Fluetsch, A., Di Lascio, E., Gerebtzoff, G., & Rodríguez-Pérez, R. (2024). Adapting Deep Learning QSPR Models to Specific Drug Discovery Projects. Molecular Pharmaceutics, 21(4), 1817-1826. https://doi.org/10.1021/acs.molpharmaceut.3c01124

Vancouver

Fluetsch A, Di Lascio E, Gerebtzoff G, Rodríguez-Pérez R. Adapting Deep Learning QSPR Models to Specific Drug Discovery Projects. Molecular Pharmaceutics. 2024;21(4):1817-1826. doi:10.1021/acs.molpharmaceut.3c01124

BibTeX
@article{Fluetsch2024,
  author = {Fluetsch, Andrin and Di Lascio, Elena and Gerebtzoff, Grégori and Rodríguez-Pérez, Raquel},
  title = {Adapting Deep Learning QSPR Models to Specific Drug Discovery Projects},
  journal = {Molecular Pharmaceutics},
  year = {2024},
  volume = {21},
  number = {4},
  pages = {1817--1826},
  doi = {10.1021/acs.molpharmaceut.3c01124}
}
Kostenlos starten →