DrugEx: Deep Learning Models and Tools for Exploration of Drug-like Chemical Space

Abstract

The discovery of novel molecules with desirable properties is a classic challenge in medicinal chemistry. With the recent advancements of machine learning, there has been a surge of de novo drug design tools. However, few resources exist that are both user-friendly as well as easily customisable. In this application note, we present the new versatile open-source software package DrugEx for multi-objective reinforcement learning. This package contains the consolidated and redesigned scripts from the prior DrugEx papers including multiple generator architectures and a variety of scoring tools and multi-objective optimisation methods. It has a flexible application programming interface and can readily be used via the command line interface or the graphical user interface GenUI. The DrugEx package is publicly available at https://github.com/CDDLeiden/DrugEx

Streamlining Computational Fragment-Based Drug Discovery through Evolutionary Optimization Informed by Ligand-Based Virtual Prescreening

“…Computational de novo drug design involves the use of techniques such as genetic algorithms, − reinforcement learning, including deep reinforcement learning, − generative deep learning models, − or other deep learning methods, e.g., graph transformers, , models that blend deep learning and evolutionary algorithms, ,, and string-based transformers (i.e., operating on a simplified molecular-input line-entry system (SMILES) string representation of molecules). , The algorithms “computationally synthesize” novel drug molecules, either by starting from scratch and adding atoms to form a novel molecule or by modifying or adding atoms to an existing chemical structure (“scaffold”). The result is the creation of novel molecules by (a) simulating chemical modifications that optimize for the single objective of improving binding efficiency to a target or (b) multiobjective optimization including drug-likeness objectives, e.g., solubility and other drug-likeness factors. ,− …”

An AI-Driven Framework for Discovery of BACE1 Inhibitors for Alzheimer’s Disease

“…18,23 Through various deep learning architectures, such as recurrent neural networks (RNNs), variational autoencoders (VAEs), and generative adversarial networks (GANs), generative AI has already generated compounds to inhibit a variety of target proteins. [24][25][26][27][28][29][30][31][32][33] Many of these compounds have displayed promising efficacy during subsequent steps of the drug design pipeline, like in silico and in vitro testing, demonstrating the effectiveness of this new drug discovery paradigm. 18 This work applies generative AI to the de novo discovery of BACE1 inhibitors.…”