Advances in De Novo Drug Design: From Conventional to Machine Learning Methods

Abstract

. De novo drug design is a computational approach that generates novel molecular structures from atomic building blocks with no a priori relationships. Conventional methods include structure-based and ligand-based design, which depend on the properties of the active site of a biological target or its known active binders, respectively. Artificial intelligence, including machine learning, is an emerging field that has positively impacted the drug discovery process. Deep reinforcement learning is a subdivision of machine learning that combines artificial neural networks with reinforcement-learning architectures. This method has successfully been employed to develop novel de novo drug design approaches using a variety of artificial networks including recurrent neural networks, convolutional neural networks, generative adversarial networks, and autoencoders. This review article summarizes advances in de novo drug design, from conventional growth algorithms to advanced machine-learning methodologies and highlights hot topics for further development.

In Silico Identification and Evaluation of Natural Products as Potential Tumor Necrosis Factor Function Inhibitors Using Advanced Enalos Asclepios KNIME Nodes

“…Towards the identification of prospective PPI compounds with TNF inhibitory action in chemical libraries datasets, our group has further optimized the Enalos computational drug discovery pipeline. Conclusively, in this paper, we used our in silico pipeline to discover new NP lead compounds as potential TNF inhibitors, continuing our previous work in the field [ 9 , 36 , 44 , 45 , 46 ].…”

Unsupervised Machine Learning Approach for Identifying Biomechanical Influences on Protein-Ligand Binding Affinity

“…Using known information on fragments such as the two discussed in this study (CCCH and CCCCCH), synthetic ligands can be chemically designed to bind optimally to a target protein [42,43]. Computational tools (including, but not limited to, ML models) can also be developed to design novel synthetic drugs using known relationships between ligand fragments [44][45][46]. Gathering a clear, data-driven understanding of ligand fragment activity is a signi cant method by which synthetic drug design for new medications can be improved.…”

Comprehensive Survey of Recent Drug Discovery Using Deep Learning

“…In general, when classifying the de novo drug designs, studies are classified based on the DL models [ 87 ]. However, in the case of actual implementation, it may be an appropriate classification; however, it may not be enough to understand the purpose of the model.…”

“…The LSTM can be used with good performance even on longer sequential data compared to the RNN. Since its introduction, various modifications of the LSTM have been proposed [ 87 ], and recently, gated recurrent units (GRU) with a simpler internal structure [ 88 ] has also been widely used. It can simply be used for the de novo drug design, which randomly generates short-length compounds, and can generate an appropriate candidate drug by inputting a target protein sequence [ 89 ].…”