HEDGEHOG: hierarchical evaluation of drug generators through rigorous filtration
Briefings in Bioinformatics · 02.11.2026 · 27 (6)
DOI: 10.1093/bib/bbag609 ↗Kurz gefasst
Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 06.11.2026. Verfasst von Daria A Ryabchenko und 7 weiteren Autor:innen. Beteiligte Einrichtung: Ligand Pro, Moscow, Russia. Hauptthemen laut MEDLINE: Drug Discovery, Software. Die Arbeit nennt 72 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Generative molecular models can support early drug discovery by proposing new candidate compounds de novo. In practice, useful candidates must balance target-relevant activity, physicochemical properties, and other multiparameter design constraints. However, standard metrics commonly used to evaluate molecular generators only weakly reflect whether the generated compounds are medicinally plausible and suitable for downstream computation. This can produce an incomplete view of model performance and inefficient use of computational resources. We introduce HEDGEHOG, a configurable six-stage evaluation framework that is constructed as a hit identification workflow: (i) preprocessing; (ii) physicochemical descriptor screening; (iii) structural alerts and graph-sanity checks; (iv) synthesis feasibility; (v) docking; and (vi) 3D pose and interaction checks. We evaluated $22$ generative models in a KRAS G12D case study, using three runs of $1000$ requested generation attempts per model. The models showed different patterns of attrition, and final survival ranged from zero to $127.0 \pm 18.1$ molecules per run. None of the standard metrics showed a significant association with final survival after correction for multiple testing. HEDGEHOG provides a reproducible framework for evaluating molecular generative models under defined screening criteria. The framework identifies stage-wise failure modes across generators and provides a practical basis for developing molecular generators better aligned with early drug discovery.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Ligand Pro, Moscow, Russia.
- Applied AI Institute, Moscow, Russia.
ORCID-Profile: Daria A Ryabchenko, Pavel Gurevich, Shamil D Kadyrov, Daria Frolova, Kseniia Fedisheva, Alexander V Shapeev, Sergei A Nikolenko, Marina A Pak
Zitieren
APA
Ryabchenko, D. A., Gurevich, P., Kadyrov, S. D., Frolova, D., Fedisheva, K., Shapeev, A. V., Nikolenko, S. A., & Pak, M. A. (2026). HEDGEHOG: hierarchical evaluation of drug generators through rigorous filtration. Briefings in Bioinformatics, 27(6). https://doi.org/10.1093/bib/bbag609
Vancouver
Ryabchenko DA, Gurevich P, Kadyrov SD, Frolova D, Fedisheva K, Shapeev AV, et al. HEDGEHOG: hierarchical evaluation of drug generators through rigorous filtration. Briefings in Bioinformatics. 2026;27(6). doi:10.1093/bib/bbag609
BibTeX
@article{Ryabchenko2026,
author = {Ryabchenko, Daria A and Gurevich, Pavel and Kadyrov, Shamil D and Frolova, Daria and Fedisheva, Kseniia and Shapeev, Alexander V and Nikolenko, Sergei A and Pak, Marina A},
title = {HEDGEHOG: hierarchical evaluation of drug generators through rigorous filtration},
journal = {Briefings in Bioinformatics},
year = {2026},
volume = {27},
number = {6},
doi = {10.1093/bib/bbag609}
}
EUVIMED