Naslov
Evaluation of Different Computer Programs for Prediction of Molecular Lipophilicity

Autori
Mornar, Ana ; Jasprica, Ivona ; Medić-Šarić, Marica ; Nikolić, Sonja

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
The 15th European Symposium on Quantitative Structure-Activity Relationships (QSAR) & Molecullar Modelling in Rational Design of Bioactive Molecules : abstracts / - Istanbul, 2004, 184-184

Skup
European Symposium on Quantitative Structure-Activity Relationships & Molecullar Modelling in Rational Design of Bioactive Molecules (15 ; 2004)

Mjesto i datum
Istanbul, Turska, 05.09.2004. - 10.09.2004

Vrsta sudjelovanja
Poster

Vrsta recenzije
Nije recenziran

Ključne riječi
evaluation; log P

Sažetak
Molecular lipophilicity (log P) is a prime physicochemical descriptor of xenobiotics with relevance to their biological properties. It plays an important role in the pharmacokinetic behavior of the drug molecules, biological activities, toxicological properties and the hydrophobic interactions of drugs with their receptors. Because experimental measurements are time consuming and difficult, computational methods are very valuble tools for prediction lipophilicity of large sets of compounds in QSPR studies, particulary at the screening stage. A number of different computer programs based on different theoretical approaches have been recently developed. The previous studies have shown better predictive power for simple organic molecules than for complex drug molecules. In our work, the predictive power of seven computer programs (HyperChem, miLogP, ALOGPs, IA logP, CLOGP, KowWin and XLOGP) was evaluated comparing obtained calculated with experimental log P values from the literature. 100 chemically and pharmacologicaly heterogeneus drug molecules with complex fuctional groups were used as a test set. Statistical analysis of results was performed using StatSoft 6.0 program. Regression analysis and Chi-Square - test were applied for evaluation of validity of the computer programs. Our analysis demostrates that the highest correlation of log P values predicted by a computer program to experimental data is attributed to KowWin program (r = 0.97 and Chi-Square = 8.19). In our further work an enlargement of the database will be done.

Izvorni jezik
Engleski

Znanstvena područja
Farmacija

POVEZANOST RADA