MolOptimizer: A Molecular Optimization Toolkit for Fragment-Based Drug Design

Abstract

MolOptimizer is a user-friendly computational toolkit designed to streamline the hit-to-lead optimization process in drug discovery. MolOptimizer extracts features and trains machine learning models using a user-provided, labeled, and small-molecule dataset to accurately predict the binding values of new small molecules that share similar scaffolds with the target in focus. Hosted on the Azure web-based server, MolOptimizer emerges as a vital resource, accelerating the discovery and development of novel drug candidates with improved binding properties.

Keywords: cheminformatics; fragment screening; hit-to-lead optimization.

Figure 1

Workflow of MolOptimizer.

Figure 2

Benchmark dataset provided with MolOptimizer. The dataset contains binding values for 2-phenylthiazole-containing drug-sized molecules that bind RNA targets. (a) A molecular structure of 2-phenylthiazole, i.e., the scaffold in each molecule, was obtained by fragment-based screening using T2 relaxation spectroscopy [5]. A representative larger molecule containing phenylthiazole was obtained using a virtual filtration approach that was applied to the ZINC database [9]. (b) Hairpin 91, located in the center of the PTC of the large ribosomal subunit of Staphylococcus aureus (PDB id. 4WCE, [10]), was the target for the virtual screening of ~800 2-phenylthiazole containing small molecules. Molecular docking was performed using Autodock 4.2 [11].

Figure 3

Example of chemical descriptors extracted using RDKit (https://www.rdkit.org/, accessed on 31 December 2023) and Mordred [12].

Figure 4

Implementation of MolOptimizer in Expert Mode (a) or Manual Mode (b). Expert Mode allows for the automatic selection of hyperparameters, whereas Manual Mode requires users to select the hyperparameters.

Figure 5

Training machine learning models with MolOptimizer.