A natural language processing approach based on embedding deep learning from heterogeneous compounds for quantitative structure-activity relationship modeling.

Over the past decade, rapid development in biological and chemical technologies such as high-throughput screening, parallel synthesis, has been significantly increased the amount of data, which requires the creation and the integration of new analytical methods, especially deep learning models. Recently, there is an increasing interest in deep learning utilization in computer-aided drug discovery due to its exceptional successful application in many fields. The present work proposed a natural language processing approach, based on embedding deep neural networks. Our method aims to transform the Simplified Molecular Input Line Entry System format into word embedding vectors to represent the semantics of compounds. These vectors are fed into supervised machine learning algorithms such as convolutional long short-term memory neural network, support vector machine, and random forest to build up quantitative structure-activity relationship models on toxicity data sets. The obtained results on toxicity data to the ciliate Tetrahymena pyriformis (IGC50 ), and acute toxicity rat data expressed as median lethal dose of treated rats (LD50 ) show that our approach can eventually be used to predict the activities of chemical compounds efficiently. All material used in this study is available online through the GitHub portal (https://github.com/BoukeliaAbdelbasset/NLPDeepQSAR.git).

Optimization of the chromone scaffold through QSAR and docking studies: Identification of potent inhibitors of ABCG2.

The membrane transporter BCRP/ABCG2 has emerged as a privileged biological target for the development of small compounds capable of abolishing multidrug resistance. In this context, the chromone skeleton was found as an excellent scaffold for the design of ABCG2 inhibitors. With the aims of optimizing and developing more potent modulators of the transporter, we herewith propose a multidisciplinary medicinal chemistry approach performed on this promising scaffold. A quantitative structure-activity relationship (QSAR) study on a series of chromone derivatives was first carried out, giving a robust model that was next applied to the design of 13 novel compounds derived from this nucleus. Two of the most active according to the model's prediction, namely compounds 22 (5-((3,5-dibromobenzyl)oxy)-N-(2-(5-methoxy-1H-indol-3-yl)ethyl)-4-oxo-4H-chromene-2-carboxamide) and 31 (5-((2,4-dibromobenzyl)oxy)-N-(2-(5-methoxy-1H-indol-3-yl)ethyl)-4-oxo-4H-chromene-2-carboxamide), were synthesized and had their biological potency evaluated by experimental assays, confirming their high inhibitory activity against ABCG2 (experimental EC50 below 0.10 µM). A supplementary docking study was then conducted on the newly designed derivatives, proposing possible binding modes of these novel molecules in the putative ligand-binding site of the transporter and explaining why the two aforementioned compounds exerted the best activity according to biological data. Results from this study are recommended as references for further research in hopes of discovering new potent inhibitors of ABCG2.