Molecular De Novo Design through Transformer-based Reinforcement Learning

TL;DR

This paper presents a Transformer-based reinforcement learning method for molecular de novo design, outperforming RNN models in generating biologically active compounds and enabling diverse applications like scaffold hopping and library expansion.

Contribution

It introduces a novel Transformer-based generative model fine-tuned with reinforcement learning for molecular design, demonstrating superior performance over RNN-based models.

Findings

Outperforms RNN models in generating active compounds

Effective in scaffold hopping and library expansion

Generates molecules with high predicted biological activity

Abstract

In this work, we introduce a method to fine-tune a Transformer-based generative model for molecular de novo design. Leveraging the superior sequence learning capacity of Transformers over Recurrent Neural Networks (RNNs), our model can generate molecular structures with desired properties effectively. In contrast to the traditional RNN-based models, our proposed method exhibits superior performance in generating compounds predicted to be active against various biological targets, capturing long-term dependencies in the molecular structure sequence. The model's efficacy is demonstrated across numerous tasks, including generating analogues to a query structure and producing compounds with particular attributes, outperforming the baseline RNN-based methods. Our approach can be used for scaffold hopping, library expansion starting from a single molecule, and generating compounds with high predicted activity against biological targets.