TY - JOUR
T1 - Application of quantitative structure-property relationship analysis to estimate the vapor pressure of pesticides
AU - Goodarzi, Mohammad
AU - dos Santos Coelho, Leandro
AU - Honarparvar, Bahareh
AU - Ortiz, Erlinda V.
AU - Duchowicz, Pablo R.
N1 - Funding Information:
We thank the financial support provided by the National Research Council of Argentina ( CONICET ) PIP11220100100151 project and to Ministerio de Ciencia, Tecnología e Innovación Productiva for the electronic library facilities. EVO and PRD are members of the scientific researcher career of CONICET.
Publisher Copyright:
© 2016 Elsevier Inc.
PY - 2016/6/1
Y1 - 2016/6/1
N2 - The application of molecular descriptors in describing Quantitative Structure Property Relationships (QSPR) for the estimation of vapor pressure (VP) of pesticides is of ongoing interest.In this study, QSPR models were developed using multiple linear regression (MLR) methods to predict the vapor pressure values of 162 pesticides. Several feature selection methods, namely the replacement method (RM), genetic algorithms (GA), stepwise regression (SR) and forward selection (FS), were used to select the most relevant molecular descriptors from a pool of variables. The optimum subset of molecular descriptors was used to build a QSPR model to estimate the vapor pressures of the selected pesticides. The Replacement Method improved the predictive ability of vapor pressures and was more reliable for the feature selection of these selected pesticides. The results provided satisfactory MLR models that had a satisfactory predictive ability, and will be important for predicting vapor pressure values for compounds with unknown values. This study may open new opportunities for designing and developing new pesticide.
AB - The application of molecular descriptors in describing Quantitative Structure Property Relationships (QSPR) for the estimation of vapor pressure (VP) of pesticides is of ongoing interest.In this study, QSPR models were developed using multiple linear regression (MLR) methods to predict the vapor pressure values of 162 pesticides. Several feature selection methods, namely the replacement method (RM), genetic algorithms (GA), stepwise regression (SR) and forward selection (FS), were used to select the most relevant molecular descriptors from a pool of variables. The optimum subset of molecular descriptors was used to build a QSPR model to estimate the vapor pressures of the selected pesticides. The Replacement Method improved the predictive ability of vapor pressures and was more reliable for the feature selection of these selected pesticides. The results provided satisfactory MLR models that had a satisfactory predictive ability, and will be important for predicting vapor pressure values for compounds with unknown values. This study may open new opportunities for designing and developing new pesticide.
KW - Multivariable Linear Regression (MLR)
KW - Pesticides
KW - Quantitative Structure-Property Relationships (QSPR)
KW - Vapor pressure (VP)
UR - http://www.scopus.com/inward/record.url?scp=84957975140&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84957975140&partnerID=8YFLogxK
U2 - 10.1016/j.ecoenv.2016.01.020
DO - 10.1016/j.ecoenv.2016.01.020
M3 - Article
C2 - 26890190
AN - SCOPUS:84957975140
SN - 0147-6513
VL - 128
SP - 52
EP - 60
JO - Ecotoxicology and Environmental Safety
JF - Ecotoxicology and Environmental Safety
ER -