Oup (qC1 ). These descriptors had been applied to establish the QSPR models by the common equation: pKa = pH qH pO qO pC1 qC1 p (two)exactly where pH , pO , pC1 , pOD , pC1D and p are parameters of your QSPR model.Descriptors and QSPR models for carboxylic acidswhere pH , pO , pC1 and p are parameters with the QSPR model (i.e., constants derived by numerous linear regression). The 5d QSPR models employ the above talked about descriptors qH , qO and qC1 and on top of that also the charge on the phenoxide O from the dissociated molecule (qOD ), as well as the charge around the carbon atom binding this oxygen (qC1D ). Applying the charges in the dissociated molecules for pKa prediction was inspired by the function of Dixon et al. [19]. The equation of your 5d QSPR models is as a result: pKa = pH H pO O pC1 C1 pOD OD pC1D C1D p (three)The descriptors were once again atomic charges and, similarly as for phenols, two kinds of QSPR models were developed and evaluated. Specifically, QSPR models with 4 descriptors (4d QSPR models) and QSPR models with seven descriptors (7d QSPR models). The 4d QSPR models used similar descriptors because the 3d models for phenols the atomic charge with the hydrogen atom in the COOH group (qH ), the charge around the hydrogen bound oxygen atom in the COOH group (qO ), as well as the charge around the carbon atom binding the COOH group (qC1 ). Additionally, also the charge from the second carboxyl oxygen (qO2 ) is incorporated. These 4d QSPR models are represented by the equation: pKa = pH qH pO qO pO2 qO2 pC1 qC1 p (four) where pH , pO , pO2 , pC1 and p are parameters in the QSPR model. The 7d QSPR models employ also charges in the dissociated forms, namely the charge around the carboxyl oxygens (qOD , qO2D ) and the charge around the carboxylic carbon atom (qC1D ). The equation on the 7d QSPR models is consequently: pKa = pH qH pO qO pO2 qO2 pC1 qC1 pOD qOD pO2D qO2D pC1D qC1D p (5)SvobodovVaekovet al. Journal of Cheminformatics 2013, five:18 a r a http://www.jcheminf.com/content/5/Page 5 ofwhere pH , pO , pO2 , pC1 , pOD , pO2D , pC1D and p are parameters from the QSPR model.QSPR model parameterizationcorrelation amongst experimental and calculated pKa are visualized in Figure two.1H,1H-Perfluoro-3,6,9-trioxadecan-1-ol Order Prediction of pKa applying EEM chargesThe parameterization in the QSPR models was carried out by multiple linear regression (MLR) applying the software program tool QSPR Designer [62].Formula of Thiol-C2-PEG2-OH Benefits and discussionQM and EEM QSPR models for phenolsWe prepared one particular 3d QSPR model and one particular 5d QSPR model making use of atomic charges calculated by every with the above talked about 18 EEM parameter sets.PMID:23927631 These models are denoted 3d or 5d EEM QSPR models. Additionally, we developed one 3d and a single 5d QSPR model using atomic charges calculated by every single of the corresponding 8 QM charge calculation approaches (denoted as 3d or 5d QM QSPR models). The information set of 74 phenol molecules was utilized for the parameterization of the QSPR models, as well as the obtained models have been validated for all molecules inside the information set. The parameterization in the 3d EEM QSPR models showed that many molecules in the data set execute as outliers. Because of this, we created also EEM QSPR models with no outliers (i.e., EEM QSPR models for which parameterization was accomplished utilizing a data set that excluded the previously observed outliers). These models are denoted 3d EEM QSPR WO models. We classified as outliers 10 from the molecules from our data set, which had the highest Cook’s square distance. Therefore the 3d EEM QSPR WO models were parameterized making use of 67 molecules, and their validation was al.