Using a Novel Consensus-Based Chemoinformatics Approach to Predict ADMET Properties and Druglikeness of Tyrosine Kinase Inhibitors.

Abstract

The urgent need to reduce the cost of new drug discovery has led us to create a new, more selective screening method using free chemoinformatics tools to restrict the high failure rates of lead compounds (>90%) during the development process because of the lack of clinical efficacy (40-50%), unmanageable toxicity (30%), and poor drug-like properties (10-15%). Our efforts focused on new molecular entities (NMEs) with reported activity as tyrosine kinase inhibitors (small molecules) as a class of great potential. The criteria for the new method are acceptable Druglikeness, desirable ADME (absorption, distribution, metabolism, and excretion), and low toxicity. After a bibliographic review, we first selected the 29 most promising compounds, always according to the literature, then collected the in silico calculated data from different platforms, and finally processed them together to conclude at 14 compounds meeting the aforementioned criteria. The novelty of the present screening method is that for the evaluation of the compounds for Druglikeness, and ADMET properties (absorption, distribution, metabolism, excretion, and toxicity), the data of the different platforms were used as a whole, rather than the results of each platform individually. Additionally, we validated our new consensus-based method by comparing the final in silico results with the experimental values of FDA (Food and Drug Administration)-approved tyrosine kinase drugs. Using inferential statistics of 39 FDA-approved tyrosine kinase drugs obtained after applying our method, we delineated the intervals of the desired values of the physicochemical properties of future active compounds. Finally, molecular docking studies enhance the credibility of the applied method as an identification tool of Druglikeness.