A quantitative structure–activity relationships study for the anti-HIV-1 activities of 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine derivatives using the multiple linear regression and partial least squares methodologies [Supplementary Material]

A quantitative structure–activity relationships (QSAR) study using Multiple Linear Regression (MLR) and Partial Least Squares (PLS) methodologies was performed for a series of 127 derivatives of 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT), a potent inhibitor of the of the human immunode...

Full description

Saved in:
Bibliographic Details
Published inJournal of the Serbian Chemical Society Vol. 78; no. 4; p. 495
Main Authors Ivan, Daniela, Crisan, Luminita, Funar-Timofei, Simona, Mracec, Mircea
Format Journal Article
LanguageEnglish
Published Belgrade Journal of the Serbian Chemical Society 01.01.2013
Subjects
Biological activity
Correlation coefficients
Dependent variables
Derivatives
Goodness of fit
HIV
Human immunodeficiency virus
Independent variables
Least squares method
Mathematical models
Regression analysis
Test sets
Thymine
Viruses
Online AccessGet full text

Cover

More Information
Summary:A quantitative structure–activity relationships (QSAR) study using Multiple Linear Regression (MLR) and Partial Least Squares (PLS) methodologies was performed for a series of 127 derivatives of 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT), a potent inhibitor of the of the human immunodeficiency virus type 1, HIV-1 reverse transcriptase (RT). The MLR and PLS methods were employed to explore the relationship between the descriptors (as independent variables) of a pool of HEPT derivative and anti-HIV-1 activity, expressed as log (1/EC50) (as dependent variables). Using Dragon descriptors, the present study was aimed at developing a predictive and robust QSAR model for predicting anti-HIV activity of HEPT derivatives for a better understanding of the molecular features of these compounds important for their biological activity. According to the squared correlation coefficients, which had values between 0.826 and 0.809 for the MLR and PLS methods, the results demonstrated almost identical qualities and good predictive ability for both the MLR and PLS models. After dividing the dataset into training and test sets, the model predictability was tested by several parAMeters, including the Golbraikh–Tropsha external criteria and the goodness of fit, tested using the Y- randomization test.
ISSN:0352-5139
1820-7421
DOI:10.2298/JSC120713085I