CHEIC: Chemical Image Classificator. An intelligent system for identification of volatiles compounds with potential for respiratory diseases using Deep Learning

Abstract

CHEIC (Chemical Image Classificator) is a web platform that uses convolutional neural networks (CNNs) to identify patterns of volatile molecules from microorganisms with potential against respiratory and bronchopulmonary diseases such as SARS-CoV-2 and asthma. The platform lets users find volatile molecules with biological activity through molecular docking. This work presents the functionalities of the CHEIC platform, which accommodates a Deep Learning model with 93% accuracy in classifying volatile compounds and candidate molecules for respiratory disease drugs. The artificial intelligence model indicated that out of 548 molecules used, 39 exhibited drug-like molecular features. By combining the indicative results of molecular docking emitted by the CHEIC platform with further analysis using 100 ns of Molecular Dynamics trajectory, four volatile compounds were found with the potential to modulate proteins associated with respiratory tract diseases such as asthma and SARS-CoV-2. Furthermore, CHEIC is a free and promising tool in the search for therapeutic agents for respiratory diseases, as well as providing valuable and fast insights for researchers interested in omics sciences.