MTMol-GPT: De novo multi-target molecular generation with transformer-based generative adversarial imitation learning
Chengwei Ai,Hongpeng Yang,Xiaoyi Liu,Ruihan Dong,Yijie Ding,Fei Guo
DOI: https://doi.org/10.1371/journal.pcbi.1012229
2024-06-27
PLoS Computational Biology
Abstract:De novo drug design is crucial in advancing drug discovery, which aims to generate new drugs with specific pharmacological properties. Recently, deep generative models have achieved inspiring progress in generating drug-like compounds. However, the models prioritize a single target drug generation for pharmacological intervention, neglecting the complicated inherent mechanisms of diseases, and influenced by multiple factors. Consequently, developing novel multi-target drugs that simultaneously target specific targets can enhance anti-tumor efficacy and address issues related to resistance mechanisms. To address this issue and inspired by Generative Pre-trained Transformers (GPT) models, we propose an upgraded GPT model with generative adversarial imitation learning for multi-target molecular generation called MTMol-GPT. The multi-target molecular generator employs a dual discriminator model using the Inverse Reinforcement Learning (IRL) method for a concurrently multi-target molecular generation. Extensive results show that MTMol-GPT generates various valid, novel, and effective multi-target molecules for various complex diseases, demonstrating robustness and generalization capability. In addition, molecular docking and pharmacophore mapping experiments demonstrate the drug-likeness properties and effectiveness of generated molecules potentially improve neuropsychiatric interventions. Furthermore, our model's generalizability is exemplified by a case study focusing on the multi-targeted drug design for breast cancer. As a broadly applicable solution for multiple targets, MTMol-GPT provides new insight into future directions to enhance potential complex disease therapeutics by generating high-quality multi-target molecules in drug discovery. Advancing drug discovery with the vast scale of possible structures of drug-like compounds (between 10 23 and 10 60 ) relies on Artificial Intelligence (AI). The process of generating SMILES has been accomplished by using language models like the Generative Pre-trained Transformer (GPT) model. In the meantime, regulating multiple targets to achieve the desired physiological responses is booming for treating complicated diseases. However, the recent deep generative models focus on single-target drug generation, and single-target drugs have limitations, such as requiring higher doses and causing mutations. For example, activation of Src kinase in certain breast cancer cells leads to EGFR phosphorylation and downstream effects. Consequently, designing multi-target molecules against multi-target simultaneously is highly required. Thus, we have proposed an upgraded GPT model MTMol-GPT with generative adversarial imitation learning for generating high-quality multiple targets drug-like compounds. For example, MTMol-GPT generates novel drugs targeting protein pairs such as DRD2 and HTR1A, EGFR and Src. As a broadly applicable tool, MTMol-GPT provides a rapid and accurate method for advancing drug discovery, ensuring the efficient generation of high-quality multi-target drugs.
biochemical research methods,mathematical & computational biology