MERMAID: an open source automated hit-to-lead method based on deep reinforcement learning

Daiki Erikawa,Masakazu Sekijima,Nobuaki Yasuo

doi:10.1186/s13321-021-00572-6

Abstract

The hit-to-lead process makes the physicochemical properties of the hit molecules that show the desired type of activity obtained in the screening assay more drug-like. Deep learning-based molecular generative models are expected to contribute to the hit-to-lead process. The simplified molecular input line entry system (SMILES), which is a string of alphanumeric characters representing the chemical structure of a molecule, is one of the most commonly used representations of molecules, and molecular generative models based on SMILES have achieved significant success. However, in contrast to molecular graphs, during the process of generation, SMILES are not considered as valid SMILES. Further, it is quite difficult to generate molecules starting from a certain molecule, thus making it difficult to apply SMILES to the hit-to-lead process. In this study, we have developed a SMILES-based generative model that can be generated starting from a certain molecule. This method generates partial SMILES and inserts it into the original SMILES using Monte Carlo Tree Search and a Recurrent Neural Network. We validated our method using a molecule dataset obtained from the ZINC database and successfully generated molecules that were both well optimized for the objectives of the quantitative estimate of drug-likeness (QED) and penalized octanol-water partition coefficient (PLogP) optimization. The source code is available at https://github.com/sekijima-lab/mermaid.

Highlights

IntroductionFrom drug discovery to launch, a new drug takes an average of 10 to 15 years to be developed and costs $2.6 billion [1, 2]
8000 drugs are currently being developed worldwide [1]
Hit-to-Lead is a stage in early drug discovery where small molecule compounds hit by high-throughput screening (HTS) are processed through certain optimizations to identify promising lead compounds [8]

Summary

Introduction

From drug discovery to launch, a new drug takes an average of 10 to 15 years to be developed and costs $2.6 billion [1, 2]. After the target protein of a therapeutic drug for a disease has been determined, high-throughput screening (HTS) is used to exhaustively test the binding affinity of thousands to hundreds of thousands of compounds to the Erikawa et al Journal of Cheminformatics (2021) 13:94. [5,6,7], and those that are determined to be active proceed to the hit-to-lead. Hit-to-Lead is a stage in early drug discovery where small molecule compounds hit by high-throughput screening (HTS) are processed through certain optimizations to identify promising lead compounds [8]. Machine learning (ML) methods such as random forests and deep learning have been used in virtual screening [9,10,11,12]; molecular design methods using generative models are expected to be used in hit-to-lead [13]

Objectives

Methods

Results

Discussion

Conclusion