Approaches to Scaffold Hopping for Identifying New Bioactive Compounds and the Contribution of Artificial Intelligence

Abstract

The failure of a potential drug in the clinical study phase represents a significant cost and a loss of time for the process. Therefore, the development of strategies that can increase efficiency, streamline efforts, and lower expenses has become one of the main goals in drug development. In this context, computer-aided drug design (CADD) and the advent of artificial intelligence (AI) have played a pivotal role in driving significant advancements in these processes. Once a bioactive compound is identified, structural optimization to achieve a pharmacokinetic profile compatible with good oral bioavailability becomes indispensable. Scaffold hopping, a time-honored strategy in medicinal chemistry, has witnessed a resurgence in popularity with the advent of artificial intelligence (AI). The combination of traditional techniques with AI models, primarily based on deep learning (DL), has increased the success of these strategies. In this review, we aim to compile articles from recent literature published in the Web of Science, PubMed, and Google Scholar databases. After removing duplicate data, we analyzed the results that identified new bioactive compounds using AI tools. The generation of pharmacophore models, molecules formed by fragment linking, ring modification, and molecular recombination, as well as compounds acting on various targets with the aid of software powered by various types of AI, has shown promising results. However, challenges persist, including issues related to the quality of input data, the interpretation and interpretability of results, regulatory matters, investments in technology, and the formation and training of multidisciplinary teams. Overall, scaffold hopping combined with AI represents a powerful approach to expedite drug discovery and facilitate the development of innovative therapeutic agents with improved efficacy and safety profiles. A general discussion of AI models in the pharmaceutical industry is also presented.

Keywords: Scaffold hopping; artificial intelligence; bioactive molecules.; deep learning; drug design; drug discovery; machine learning.