DN-Based DTI Model to Identify Potential Drug Molecules Against COVID-19
Santhosh Amilpur, Chandra Mohan Dasari · 2023
The novel coronavirus COVID-19 has created havoc in each of our lives, causing uncertainty, psychological stress, health anxiety, and financial stress. There are no approved antiviral drugs to combat coronaviruses. A deep learning technique for drug-target interaction (DTI) has recently emerged as an innovative field of research. Because of the extremely complex drug-target structure, extracting relevant information to anticipate targets has proven difficult. In response, there has been a large-scale experimental and computational research effort to study and develop drugs. Deep learning algorithms are applied in this process to produce novel drug candidates that have the potential to be effective at searching through a wide range of molecules. A promising area in the field of drug development has recently emerged due to advances in reinforcement learning combined with generative methods. In this chapter, a two-step approach is proposed for identifying potential drug molecules to combat COVID-19 and its variants. Initially, by inducing generative adversarial networks (GAN) using reinforcement techniques, we generate novel molecules. Then, fine-tuning the model with existing repurposed molecules that are active towards the target protease, we identify a finite set of molecules that bind to the main protease of SARS-CoV-2. The generative model&s;s ability is validated by determining various similarity metrics for the generated novel molecules. Later, to determine the binding affinity between potential compounds and the target protease sequence, we suggest a deep learning-based unique drug-target interaction (DTI) model, in contrast to earlier approaches that depend on docking processes. Finally, the binding affinity of the generated molecules is predicted against the 3CLpro main protease by using the proposed DTI model.