The Evolution of Machine Learning in Medicinal Chemistry: A Comprehensive Bibliometric Analysis

In the medicinal chemistry (MC) field, artificial intelligence (AI) has been used to establish quantitative structure-activity relationship (QSAR) classification models, virtual screening, drug discovery, drug design, and so on. In this investigation, MC AI studies (AI-MC) (from 2001-2023) underwent quantitative and qualitative modeling analyses.

Using a hybrid research strategy incorporating content analyses and bibliometric methods, we retrospectively analysed the AI-MC literature using a bibliometrix package (R software) combined with CiteSpace V and VOSviewer programs.

Between 2001 and 2023, AI-MC articles were published in 92 countries or regions, with China and the United States leading in the number of publications. Also, 196 affiliations were added to AI-MC research; the CHINESE ACADEMY OF SCIENCES contributed the most. Reference clusters were categorized as follows: (1) QSAR, (2) virtual screening, (3) drug discovery, (4) drug design. Predictive model (2020-2021), molecular fingerprints (2021-2023) and scoring function (2021-2023) reflected research frontier keywords. As we look to the future, the ongoing progress and innovation in technology herald the promising development of multimodal and large language models (LLMs) within the realm of MC.

The integration of AI into MC has significantly transformed the landscape of drug development. AI techniques, particularly machine learning, and deep learning algorithms, have demonstrated remarkable potential in accelerating the discovery and optimization of new drugs. By leveraging large datasets and advanced computational models, AI enhances the efficiency of virtual screening, improves the accuracy of QSAR models, and facilitates the design of novel therapeutic agents. As the technology continues to advance, the development of multimodal and large language models (LLMs) is expected to further revolutionize this field, offering new opportunities for more precise and efficient drug design and discovery.

We comprehensively characterized the AI-MC field and determined future trends and hotspots. Importantly, we provided a dynamic oversight of the AI-MC literature and identified key upcoming research areas.