Intelligence computational analysis of letrozole solubility in supercritical solvent via machine learning models.

Autor: Alqarni M; Department of Pharmaceutical Chemistry, College of Pharmacy, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia. m.aalqarni@tu.edu.sa., Ashour AA; Department of Oral & Maxillofacial Surgery and Diagnostic Sciences, Faculty of Dentistry, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia., Shafie A; Department of Clinical Laboratory Sciences, College of Applied Medical Sciences, Taif University, P.O.Box 11099, Taif, 21944, Saudi Arabia., Alqarni A; Department of Oral & Maxillofacial Surgery and Diagnostic Sciences, Faculty of Dentistry, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia., Felemban MF; Department of Oral & Maxillofacial Surgery and Diagnostic Sciences, Faculty of Dentistry, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia., Shukr BS; Department of Preventive Dentistry, Faculty of Dentistry, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia., Alzubaidi MA; Department of Preventive Dentistry, Faculty of Dentistry, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia., Algahtani FS; Department of Restorative Dental Science, Faculty of Dentistry, Taif University, Taif, 21944, Saudi Arabia.
Jazyk: angličtina
Zdroj: Scientific reports [Sci Rep] 2024 Sep 17; Vol. 14 (1), pp. 21677. Date of Electronic Publication: 2024 Sep 17.
DOI: 10.1038/s41598-024-73029-z
Abstrakt: Supercritical fluids (SCFs) can be used to prepare drugs nanoparticles with improved solubility. SCFs have shown superior advantages in pharmaceutical industry as an environmentally friendly alternative to toxic/harmful organic solvents. They possess gas-like transport characteristics and liquid-like solvation power for solutes. Evaluation of chemotherapeutic drugs' solubility in supercritical carbon dioxide (SCCO 2 ) has been recently an attractive subject for developing this method in pharmaceutical sector. To reach this purpose, the utilization of accurate models is of great necessity to estimate experimental-based solubility data. In this paper, the authors tried to employ machine learning (ML) approaches to estimate the solubility of Letrozole (LET) drug as chemotherapeutic agent and correlate its values in wide ranges of temperature and pressure. To do this, PAR (Passive Aggressive Regression), RF (Random Forest), and RBF-SVM are the models used (Support Vector Machine with RBF kernel). These models optimized in terms of their hyper-parameters using GA algorithm. The optimized PAR, RF, RBF-SVM models obtained coefficients of determination (R-squared) of 0.8277, 0.9534, and 0.9947. Also, the MSE error rate of the models are 0.1342, 0.0305, and 0.0045, in the same order. The final result of the evaluations shows the optimized RBF-SVM model as the most appropriate model in this research. The model exhibits a maximum prediction error of 0.1289.
(© 2024. The Author(s).)
Databáze: MEDLINE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje