Transformative drug research using augmented generative AI

Abstract

The conventional method of finding new drugs is time-consuming, costly, and frequently wasteful, posing a challenge to pharmaceutical firms. Generative Artificial Intelligence (AI) provides a better solution by quickly producing candidate drug molecules from patterns learned from large chemical databases. All of these processes are guided by a retrieval-augmented-generation system that incorporates relevant biomedical facts to guide molecular generation, ensuring alignment with known biological mechanisms. Furthermore, AI models find great utility in the generation and manipulation of molecular structures with standard representations. Once a compound is generated, it is screened against different drug-likeness parameters for acceptability in humans. In addition, advanced Natural Language Processing methods help extract important information from scientific articles, patents, and clinical reports, refining candidate selection and design.

This integration of artificial intelligence facilitates an even more systematic and focused approach to drug discovery that allows researchers to rapidly identify potential compounds with fewer resources. Through machine learning algorithms, scientists are predicting molecular interactions, designing drug properties, and recognizing potentially hazardous toxicity in the most effective manner. This not only speeds up drug discovery but also increases the chances of trial success. Artificial intelligence-assisted drug discovery is the lifeblood of reforms in pharmaceutical research that mark the beginning of faster, cheaper, and innovative cures for many diseases.

Keywords

Artificial Intelligence machine learning drug delivery Large Language Model Natural Language Processing retrieval – augmented generation simplified molecular input line entry system Generative Pretrained Transformer

Get full access to this article

View all access options for this article.

References

Deng

Yang

Ojima

, et al. Artificial intelligence in drug discovery: applications and techniques. https://doi.org/10.48550/arXiv.2106.05386.

Wang

Zhao

, et al. A transformer-based generative model for de novo molecular design. https://doi.org/10.48550/arXiv.2210.08749.

Nguyen

Tran

. Learning to discover medicines. https://doi.org/10.48550/arXiv.2202.07096.

Filella-Merce

Molina

Orzechowski

, et al. Optimizing drug design by merging generative AI with active learning frameworks. https://doi.org/10.48550/arXiv.2305.06334.

Frey

Soklaski

Axelrod

, et al. Neural scaling of deep chemical models. ChemRxiv, 2022. doi: 10.26434/chemrxiv-2022-3s51.

Kong

Huang

Tan

, et al. Molecule generation by principal subgraph mining and assembling. https://doi.org/10.48550/arXiv.2106.15098.

Honda

Shi

Ueda

. SMILES transformer: pre-trained molecular fingerprint for low data drug discovery. https://doi.org/10.48550/arXiv.1911.04738.

Lewis

Perez

Piktus

, et al. Retrieval-augmented generation for knowledge-intensive NLP tasks. https://doi.org/10.48550/arxiv.2005.11431.

siriwardhana

weerasekera

wen

, et al. fine-tune the entire RAG Architecture. https://doi.org/10.48550/arXiv.2106.11517.

10.

Asai

Wang

, et al. SELF-RAG: learning to retrieve, generate, and critique through self-reflection. https://doi.org/10.48550/arXiv.2310.11511.

11.

Zhao

Chu

. AutoML: a survey of the state-of-the-art. Knowl-Based Sys 2021; 212: 106622.

12.

Burke

Gendreau

Hyde

, et al. Hyper-heuristics: a survey of the state of the art. J Oper Res Soc 2013; 64: 1695–1724.

13.

The role of AI in drug discovery: challenges, opportunities, and transforming the pharmaceutical landscape. https://doi.org/10.3390/ijms241410000.

14.

Reinvent 4: modern AI–driven generative molecule design. https://doi.org/10.1186/s13321-024-00812-5.

15.

Artificial intelligence in drug discovery and development. https://doi.org/10.1016/j.drudis.2020.09.014.

16.

Malusare

Aggarwal

. Improving molecule generation and drug discovery with a knowledge-enhanced generative model. https://doi.org/10.48550/arXiv.2402.08790.

17.

Yin

Yang

, et al. An open unified deep graph learning framework for discovering drug leads. https://doi.org/10.48550/arXiv.2301.03424.

18.

Kadan

Ryczko

Roitberg

, et al. Guided multi-objective generative AI to enhance structure-based drug design. https://doi.org/10.48550/arXiv.2405.11785.

19.

Generative AI in drug discovery and development. https://doi.org/10.1016/j.drudis.2024.09.014.

20.

Generative artificial intelligence in drug discovery: basic framework and applications. https://doi.org/10.3389/fphar.2024.1331062.

21.

Unleashing the power of generative AI in drug discovery. https://doi.org/10.1016/j.drudis.2024.10.005.

22.

Generative AI develops potential new drugs for antibiotic-resistant bacteria. https://med.stanford.edu/news/all-news/2024/03/ai-drug-development.html.

23.

Generative AI in the pharmaceutical industry: moving from hype to reality. https://www.mckinsey.com/industries/life-sciences/our-insights/generative-ai-in-the-pharmaceutical-industry-moving-from-hype-to-reality.

24.

NVIDIA generative AI Is opening the next era of drug discovery and design. https://blogs.nvidia.com/blog/drug-discovery-bionemo-generative-ai/.

25.

Panagoulias

Virvou

Tsihrintzis

. Augmenting large language models with rules for enhanced domain-specific interactions: the case of medical diagnosis. Electronics (Basel) 2024; 13: 20.

26.

KEGG Pathway Dataset. https://www.genome.jp/kegg/pathway.html (accessed on 23 Feb 2025).

27.

Pubmed Dataset. https://pubmed.ncbi.nlm.nih.gov/ (accessed on 23 Feb 2025).