Premium
Partition Coefficient Prediction of a Large Set of Various Drugs and Poisons by a Genetic Algorithm and Artificial Neural Network
Author(s) -
Riahi Siavash,
Beheshti Abolghasem,
Mohammadi Ali,
Ganjali Mohammad Reza,
Norouzi Parviz
Publication year - 2008
Publication title -
journal of the chinese chemical society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.329
H-Index - 45
eISSN - 2192-6549
pISSN - 0009-4536
DOI - 10.1002/jccs.200800051
Subject(s) - quantitative structure–activity relationship , artificial neural network , partition coefficient , chemistry , molecular descriptor , biological system , genetic algorithm , branching (polymer chemistry) , partition (number theory) , algorithm , artificial intelligence , computer science , machine learning , mathematics , chromatography , stereochemistry , organic chemistry , combinatorics , biology
A quantitative structure‐property relationship (QSPR) was developed using Artificial Neural Network (ANN) modeling to study partition coefficients (LogP) of various drugs and poisons. A set of 197 drugs and poisons were selected and suitable sets of molecular descriptors were calculated. A genetic algorithm was used to select important molecular descriptors and a supervised ANN was applied to correlate the selected descriptors with the experimental values of LogP. The developed model indicates that molecular branching and mass, polar parts and volume of molecule can be used to predict the partition coefficient of drugs and poisons. The described model does not require experimental parameters and potentially provides useful prediction for LogP of new drugs and poisons. For genetic algorithm neural network methods, the standard errors of calibration, test and prediction were 2.94, 2.84 and 4.38%, respectively. The predicted LogP values were compared with those obtained previously via other models on the same subject and the comparison clearly revealed the superiority of the proposed model over other models.