Quantitative Structure-Retention Relationships Studies of Selected Groups of Compounds Characterized by Different Pharmacological Activity Using Multiple Linear Regression Procedure
Jolanta Stasiak, Marcin Koba and Tomasz Baczek
Affiliation: Department of Toxicology, Faculty of Pharmacy, Nicolaus Copernicus University, Collegium Medicum of Bydgoszcz, Dr. A. Jurasza 2, 85-094 Bydgoszcz, Poland.
Keywords: HPLC, Molecular modeling, MLR, QSRR models, Retention parameters.
In this paper the quantitative structure-retention relationships (QSRR) studies for three different groups of
drugs (cardiovascular system drugs, analgesic drugs and some compounds characterized by divergent pharmacological activity)
and for chromatographic parameters (log k' or log k'
w retention factors) determined with the use of the HPLC methods
were performed. Molecular descriptors (over 4900 molecular modeling parameters) obtained using the HyperChem
and the Dragon computer programs, and the Virtual Computational Chemistry Laboratory (VCCLAB) website were applied
to derive the QSRR equations by means of multiple linear regression method with stepwise procedure. Several statistically
significant QSRR equations as two-parameters for both cardiovascular and analgesic drugs, and sevenparameters
for group of “other” drugs were built. Six QSRR equations with correlations R ranged from 0.9822 to 0.9957
were obtained for cardiovascular drugs, twenty-six QSRRs with R equal to 0.9120-0.9776 were derived for analgesics,
and ten QSRRs with R from 0.9529 to 0.9924 were calculated for “other“ drugs. Moreover, the proposed QSRR models
give important information about physico-chemical properties of analyzed drugs, and indicated that descriptors characterized
topology, geometry and lipophilicity of molecular structures of analyzed compounds are crucial for prediction of
their retention parameters. Additionally, derived QSRR models can be helpful to search (to prediction) HPLC retention
factor for the new drug candidates.
Rights & PermissionsPrintExport