Bridging chemical space and biological efficacy: advances and challenges in applying generative models in structural modification of natural products

Abstract

Natural products (NPs) are invaluable resources for drug discovery, characterized by their intricate scaffolds and diverse bioactivities. AI drug discovery & design (AIDD) has emerged as a transformative approach for the rational structural modification of NPs. This review examines a variety of molecular generation models since 2020, focusing on their potential applications in two primary scenarios of NPs structure modification: modifications when the target is identified and when it remains unidentified. Most of the molecular generative models discussed herein are open-source, and their applicability across different domains and technical feasibility have been evaluated. This evaluation was accomplished by integrating a limited number of research cases and successful practices observed in the molecular optimization of synthetic compounds. Furthermore, the challenges and prospects of employing molecular generation modeling for the structural modification of NPs are discussed.

Keywords: Artificial intelligence; Molecular generative models; Natural products; Structural modification.

Fig. 1

DeepFrag workflow: The “Parent” and “Fragment” of the initial ligand are highlighted in orange and yellow respectively, which are converted into 3D voxel grids (density channel), merged and fed into the DeepFrag model to predict the missing fragment fingerprints. Finally, the generated fragment fingerprints were compared with a predefined fingerprint library to obtain the prediction results.

Fig. 2

Two representative data-driven methods. A MolGPT illustration and B EMPIRE illustration

Fig. 3

Two representative scaffold hopping methods. A DeepHop illustration and B SyntaLinker illustration

Fig. 4

Two application examples of Deepfrag. A Antiviral drug development and B Topo IIα inhibitors optimization

Fig. 5

Two application examples of Scaffold Decorator. A Discovery of A_2B receptors and B DDR1 inhibitors

Fig. 6

A Scaffold hopping of JAK1 inhibitors and B linker optimization of TBK1 inhibitors