Guided Multi-objective Generative AI to Enhance Structure-based Drug Design DOI Creative Commons
Amit Kadan, Kevin Ryczko, Erika Lloyd

и другие.

Chemical Science, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Generative AI has the potential to revolutionize drug discovery. Yet, despite recent advances in deep learning, existing models cannot generate molecules that satisfy all desired physicochemical properties. Herein, we describe IDOLpro, a novel generative chemistry combining diffusion with multi-objective optimization for structure-based design. Differentiable scoring functions guide latent variables of model explore uncharted chemical space and ligands silico, optimizing plurality target We demonstrate our platform's effectiveness by generating optimized binding affinity synthetic accessibility on two benchmark sets. IDOLpro produces affinities over 10-20% higher than next best state-of-the-art method each test set, producing more drug-like generally better scores other methods. do head-to-head comparison against an exhaustive virtual screen large database molecules. show can range important disease-related targets any molecule found while being 100× faster less expensive run. On set experimental complexes, is first produce experimentally observed ligands. accommodate (e.g. ADME-Tox) accelerate hit-finding, hit-to-lead, lead

Язык: Английский

On the application of artificial intelligence in virtual screening DOI
Thanawat Thaingtamtanha, R Ravichandran, Francesco Gentile

и другие.

Expert Opinion on Drug Discovery, Год журнала: 2025, Номер unknown

Опубликована: Май 19, 2025

Artificial intelligence (AI) has emerged as a transformative tool in drug discovery, particularly virtual screening (VS), which is crucial initial step identifying potential candidates. This article highlights the significance of AI revolutionizing both ligand-based (LBVS) and structure-based (SBVS) approaches, streamlining enhancing discovery process. The authors provide an overview applications with focus on LBVS SBVS approaches utilized prospective cases where new bioactive molecules were identified experimentally validated. Discussion includes use quantitative structure-activity relationship (QSAR) modeling for LBVS, well its role techniques such molecular docking dynamics simulations. based literature searches all studies published up to March 2025. rapidly transforming VS by leveraging increasing amounts experimental data expanding scalability. These innovations promise enhance efficiency precision across yet challenges curation, rigorous validation models, efficient integration methods remain critical realizing AI's full discovery.

Язык: Английский

Процитировано

0
Guided Multi-objective Generative AI to Enhance Structure-based Drug Design DOI Creative Commons
Amit Kadan, Kevin Ryczko, Erika Lloyd

и другие.

Chemical Science, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Generative AI has the potential to revolutionize drug discovery. Yet, despite recent advances in deep learning, existing models cannot generate molecules that satisfy all desired physicochemical properties. Herein, we describe IDOLpro, a novel generative chemistry combining diffusion with multi-objective optimization for structure-based design. Differentiable scoring functions guide latent variables of model explore uncharted chemical space and ligands silico, optimizing plurality target We demonstrate our platform's effectiveness by generating optimized binding affinity synthetic accessibility on two benchmark sets. IDOLpro produces affinities over 10-20% higher than next best state-of-the-art method each test set, producing more drug-like generally better scores other methods. do head-to-head comparison against an exhaustive virtual screen large database molecules. show can range important disease-related targets any molecule found while being 100× faster less expensive run. On set experimental complexes, is first produce experimentally observed ligands. accommodate (e.g. ADME-Tox) accelerate hit-finding, hit-to-lead, lead

Язык: Английский

Процитировано

0