ABSTRACT
Cheminformatics plays a major part during the drug discovery processes in identifying the lead molecules that have substantial action against a selected biological target. The lead molecules can modulate the biological targets which are mostly proteins and whose modulation may bring about a positive change in controlling or curing a disease. Molecular modeling and machine learning approaches facilitate the screening and optimization of compounds against a desired target in the drug discovery process which were finally evaluated in human clinical trials. Machine learning techniques are a set of methods/approaches that are implemented to process/mine chemical databases/datasets for a range of different applications such as small molecule design, drug target identification, drug screening, etc. With the advent of deep learning neural networks, new innovative ways to mine huge chemical space have been unlocked. Machine learning approaches offer speed and facilitate the reduction of chemical space for better decision making. The potential application of molecular modeling includes the prediction of structures related with the disease progression, identifying the existing drugs which have the potency to be useful in the targeting proteins and proposing new chemical compounds useful against the disease progression. High-throughput virtual screening, pharmacophore modeling, molecular docking, and molecular dynamics simulation are essential tools in molecular modeling, which are widely used in small molecule drug screening.
