Tianyu Chen1, Tianqiang Wu1, Ningxi Li1, Yichun Jiang1, Huanli Yin1, Min Wu2. 1. Department of Pharmacy, Chengdu Medical College, Chengdu, P.R. China. 2. Department of Pharmacy, Chengdu Medical College, Chengdu, P.R. China;, Email: wuminzhaofeng@126.com.
Abstract
Background: The Biopharmaceutics Classification System (BCS), which classifies bioactive molecules based on solubility and permeability, is widely used to guide new drug development and drug formulation, as well as predict pharmacokinetics. Here we performed computer simulations to study correlations between a molecule's structure and its BCS classification. Methods: A total of 411 small molecules were assigned to BCS categories based on published drug data, and their Pybel-FP4 fingerprints were extrapolated. The information gain (IG) of each fingerprint was calculated and its characteristic structure analyzed. IG was calculated using multiple thresholds, and results were verified using support vector machine prediction, while taking into account the dose coefficient (0-0.1, 0.1-1, or >1). Structural functional features common to fingerprints of compounds in each type of BCS class were determined using computer simulations. Results: BCS classes III and IV appear to share several structural and functional characteristics, including secondary aliphaticamine, Michael acceptor, isothiourea, and sulfonamide sulfonic derivatives. Conclusion: We demonstrate that our approach can correlate characteristic fingerprints of small-molecule drugs with BCS classifications, which may help guide the development and optimization of new drugs.
Background: The Biopharmaceutics Classification System (BCS), which classifies bioactive molecules based on solubility and permeability, is widely used to guide new drug development and drug formulation, as well as predict pharmacokinetics. Here we performed computer simulations to study correlations between a molecule's structure and its BCS classification. Methods: A total of 411 small molecules were assigned to BCS categories based on published drug data, and their Pybel-FP4 fingerprints were extrapolated. The information gain (IG) of each fingerprint was calculated and its characteristic structure analyzed. IG was calculated using multiple thresholds, and results were verified using support vector machine prediction, while taking into account the dose coefficient (0-0.1, 0.1-1, or >1). Structural functional features common to fingerprints of compounds in each type of BCS class were determined using computer simulations. Results: BCS classes III and IV appear to share several structural and functional characteristics, including secondary aliphaticamine, Michael acceptor, isothiourea, and sulfonamide sulfonic derivatives. Conclusion: We demonstrate that our approach can correlate characteristic fingerprints of small-molecule drugs with BCS classifications, which may help guide the development and optimization of new drugs.
Authors: Barkat Ali Khan; Hina Khalid; Muhammad Khalid Khan; Khaled M Hosny; Shahzeb Khan; Waleed Y Rizg; Awaji Y Safhi; Abdulrahman A Halwani; Alshaimaa M Almehmady; Farid Menaa Journal: Polymers (Basel) Date: 2022-08-25 Impact factor: 4.967