Current Trends in Quantitative Structure–Activity Relationship Validation and Applications On Drug Discovery

Abstract

Quantitative structure-activity relationship history, current status & the importance of validationMethods to correlate biological activity and chemical structure of compounds have been employed since 1868 [1]. Early on, quantitative structure-activity relationship (QSAR) analyses were performed using experimentally determined physicochemical properties, such as logarithm of water/n-octanol partition coefficient (log P), hydrophobic constant (π) and Hammet electronic constant (σ), which were then correlated with the biological activity of the tested compounds [2].Nowadays, there are several prerequisites to construct and apply QSAR models. From the applicability point of view, QSAR requires a compound set that has been tested against an identified molecular target, cell, tissue, or even microorganism, under the same experimental conditions, and possesses the minimum variance in the observed responses [3]. Once an appropriate dataset has been selected, the main steps of a QSAR modeling require molecular/physicochemical properties, followed by variable selection, model generation from different algorithms and, most importantly, a validation process using internal and external datasets.After the generation and validation of a QSAR model, the model can be employed in predictions of biological activity of new samples and a physicochemical interpretation of the observed phenomena could be also conducted, providing insights for the design of new bioactive chemicals and/or their molecular mechanism of action. Subsequently, QSAR models have been widely employed in several steps of the drug design process, for instance, in order to understand and predict the compound binding affinity for a specific molecular target. QSAR modeling is also applied to better comprehend general phenomena such as pharmacokinetics and toxicity-related end points, which generally are standard measurements and for which there are available large datasets [4].Complementarily, although animal testing is still considered crucial to the evaluation of chemical safety, toxicity testing is moving toward a greater understanding of the disease at multiple biological levels, so as to develop alternative methods [5,6]. From the biological point of view, various models representing several layers of human physiology and metabolism, both healthy and disease based, have been developed. In order to address general problems, such as skin sensitization, hepatotoxicity and DNA harming agents, some research groups have built local QSAR models using mechanistic information [6,7]. However, in the future, we can easily expect that improvements in the validation techniques and data quality will follow the growth in the data generation, leading to broader QSAR models.There are several relevant articles in the QSAR field describing methods and validation techniques [8], besides extensive literature on troubleshooting, which makes this field still highly attractive for research and

Navigating bioactivity space in anti-tubercular drug discovery through the deployment of advanced machine learning models and cheminformatics tools: a molecular modeling based retrospective study

“…Graphic processing units, Cloud technology, and servers are very known and have been used in computer-aided drug design and discovery streams ( Aleksandrov and Myllykallio, 2019 ) . The system pharmacology method has gained popularity among scientists due to its ability to conduct pharmacodynamic assessments, discover new targets, and provide a systems-level understanding of drug-disease interactions ( Maltarollo et al, 2017 ) . Combining QSAR with machine learning methods enhances prediction power and is needed for the future development of QSAR ( Sabitov et al, 2017 ) .…”

Transfer and Multi-task Learning in QSAR Modeling: Advances and Challenges

“…Nowadays one can observe increasing number of applications of transfer and multi-task learning in medicinal studies. There are also current challenges in the QSAR field that comprise the integration of different datasets (even from different experiments) aiming the same or similar endpoints ( Maltarollo et al, 2017 ) and the development of universal QSAR models using very large datasets ( Alves et al, 2017 ). Therefore, good examples of dataset that could be benefited from transfer and multi-task learning are: (i) compounds with same endpoint measured under different experimental conditions; (ii) antimicrobial activities against genetically similar microorganisms; (iii) compounds with the same mechanism of action in homologous targets and high degree of similarity in the binding pocket; (iv) non-specific endpoints as toxicity against a cell line or permeability rates determined by different models.…”

Recent Advances in the Prediction of Pharmacokinetics Properties in Drug Design Studies: A Review

“…As more information regarding a specific biological process as the absorption, for example, as more accurate the future predictions will be, even with differences in experimental protocols. Nowadays, there are available techniques such as multi‐target modeling and transfer learning that could be useful to predict bioavailability using experimental data from in vivo , cell‐based, and cell‐free assays [112] …”