A New Group Contribution Approach to the Calculation of LogP

International Journal of Pharmaceutics, 2005

The experimental octanol/water partition coefficient data, of 108 compounds from the data set [Rytting, E., Lentz, K.A., Chen, X., Qian, F., Venkatesh, S., 2004. A quantitative structure-property relationship for predicting drug solubility in PEG 400/water cosolvent systems. Pharm. Res. 21, 237-244] was compared to calculated values using the computer programs ClogP, ACD/logPdb and KowWin. It was found that all the three programs have a user friendly interface but ClogP appears to be the more accurate predictor of log K(ow).

Proceedings of MOL2NET 2019, International Conference on Multidisciplinary Sciences, 5th edition

The distribution coefficient (log P) is an important molecular characteristic that allows us to estimate the lipophilicity of chemical compounds and predict how a drug will behave, fundamentally against the processes of absorption and excretion. The experimental determination of this and other properties of interest has several limitations, such as the high time invested and the consumption of considerable amounts of sample. In recent years, the development of new drugs has been supported by computational tools that allow a theoretical prediction of their properties from the information collected by their molecular descriptors, their design being much faster and cheaper. This paper shows the results of a structure-property relationship (QSPR) study aimed at finding a predictive mathematical model of the distribution coefficient of organic compounds of pharmaceutical interest. Through the computer programs ACDLabs (simplified molecular representations and calculation of log P) and MODESLAB (calculation of molecular descriptors) a training series consisting of 200 compounds classified in ten pharmacological groups was formed. Using the BuildQSAR computer program, an optimal prediction model of log P was obtained, considering the five molecular descriptors that best correlated with this property as independent variables. The model obtained showed a percentage of adjustment to the experimental data of 85%, as well as a standard error of the estimate lower than the logarithmic unit. Its internal validation showed an adjustment percentage of 80%.

Journal of Food Engineering, 2004

Risk of molecular migration in food/packaging system is important consideration from safety, hygienic and economic points of view. Octanol/water system is a good reference for explanation of hydrophobic/hydrophilic character of food/packaging system.

Chemosphere, 2007

A regression method was developed for the hydrophobicity ruler approach, which is an indirect method for determining the octanol/ water partition coefficients of very hydrophobic compounds. Two constants introduced into the mathematical model were obtained by regression of the absorption data sampled before the partition equilibrium. A water miscible organic solvent was used to increase the solubility of the very hydrophobic compounds in the aqueous solution so that the hydrophobicity scale was reduced and the equilibration was accelerated. Polydimethylsiloxane/methanol aqueous solution and a series of 21 polychlorinated biphenyls (PCBs) were used to demonstrate the regression method. The PCB compounds with known experimental log K o/w values served as reference compounds, while the PCB compounds without known log K o/w values were determined. The distribution coefficients (log K p/s ), uptake and elimination rate constants were obtained from the two regression constants for each compound (reference or unknown). The correlation of the log K p/s values of the reference PCB compounds with their log K o/w values was linear (log K o/w = 2.69 log K p/s + 0.76, R 2 = 0.97). The log K o/w values were compared with literature values and suggested that some values from the literature far off the calibration line could be inaccurate. The critical experimental factors, the merits of the regression method were discussed.

Environmental science & …, 1999

Numerous correlations have been developed between the organic carbon/water partition coefficient K OC and various molecular properties and descriptors, but most notably the octanol/water partition coefficient K OW and water solubility. From an analysis of the ...

The literature data for the values of the octanol/water partition coefficient K OW are examined critically, specifically in relation to the methylene group increment in log K OW -here called the lipicity L -for homologous series. With simply substituted alkanes, the plots of L versus the methylene group number m are linear, following the form: L = ˛ + ˇm (Collander equation). The slope parameter ˇ represents the methylene group increment, which widely is expected to be constant on the simplest theoretical grounds, and assumed to be so in most practical applications. The Collander equation behaviour for some 84 homologous series and subseries, ranging in complexity from the alkanes up to the alkyladenines and alkyl galactosides, is presented graphically. Compounds with ␣,-disubstituted alkyl chains give nonlinear Collander plots. The remaining series give linear Collander plots, but the methylene group incremenť is not constant, the variation being statistically significant, with the distribution essentially normal (Gaussian) and with the mean value¯ = 0.52 and standard deviation (ˇ) = 0.06. The literature data from other solvent/water systems -ethoxyethane (diethyl ether), and the two alkanes heptane and hexadecane -show similar behaviour. Most significantly, the fact that the methylene group increment ǐ s not constant casts doubts on the applicability of the linear free energy approach, and of the "fragmental methods" that are widely used in interpreting and predicting partition coefficients. More generally, the graphical approach used is essential in a proper treatment of correlations of this kind; the graphs form an atlas that shows at a glance the partition coefficient behaviour for these series, revealing anomalies in the literature data that need to be rectified, and gaps that need to be filled. The symbols and abbreviations used in this paper are listed in the Nomenclature section at the end. For numerical values, a set of data for a variable x with (for example) mean valuex = 1.23 and standard deviation (x) = 0.04 is summarised in the standard formx = 1.23(4).

Journal of Structural Chemistry, 2012

An artificial neural network (ANN) is constructed and trained for the prediction of gas to water partition coefficients of various organic compounds. The inputs of this neural network are theoretically derived from molecular descriptors that were chosen by the genetic algorithmpartial least squares (GA-PLS) feature selection technique. These descriptors are: areaweighted surface charge of hydrogen bonding donor atoms (HDCA-2), average bond order of a C atom (P C), Kier flexibility index ()), atomic charge weighted partial positively charged surface area (PPSA-3), and difference between atomic charge weighted partial positive and negative surface areas (DPSA-3). By comparing the results obtained from PLS and ANN models, one can see that statistical parameters (Fisher ratio, correlation coefficient, and standard error) of the ANN model are better than those of the PLS model, which indicates that a nonlinear model can simulate more accurately the relationship between the structural descriptors and the partition coefficients of the investigated molecules. K e y w o r d s: artificial neural network, gas to water partition coefficient, genetic algorithm, partial least squares.

Asian Journal of Green Chemistry, 2017

Life and its extraction fuels climate change. We performed studies upon an extended series of petroleum hydrocarbons, with octanol-water partition coefficients (log Kow), by using the quantitative structure-activity relationship (QSAR) methods that imply analysis of correlations and representation of models. A suitable set of molecular descriptors was calculated and the genetic algorithm (GA) was employed to select those descriptors, resulting in the best-fit models. The partial least squares PLS (PLS) was utilized to construct the linear QSAR model. The best GA-PLS model contains 27 selected descriptors in 10 latent variables space. The R2 and RMSE for training and test sets were (0.827, 0.088) and (0.716, 0.185), respectively. Inspection of the results reveals a higher R2 and lowers the RMSE value parameter for the data set GA-PLS. The GA-PLS linear model has good statistical quality with low prediction error. This is the first research on the QSAR which uses GA-PLS for the presiction octanol-water partition coefficients of some of the environmental toxic of the petroleum substances.

The computer-aided drug design is an important tool in modern medicinal chemistry. Molecular lipophilicity, usually quantified as log P, is an important molecular characteristic in medicinal chemistry and also in rationalized drug design. The log P coefficient is well-known as one of the principal parameters for the estimation of lipophilicity of chemical compounds and determines their pharmacokinetic properties. This parameter has been measured using known experimental methods, but recently huge progress in determination of log P using computational chemistry methods is observed. The number of methodological publications about lipophilicity predictions has gradually increased over the last ten years, but the number of programs available for an on-line prediction of this important parameter remains limited. This paper presents some of log P prediction methods and very popular programs connected to this topic. The prediction of log P is highly important for the pharmaceutical industry since it limits time-consuming experiments to measure log P required to optimize pharmacodynamic and pharmacokinetic properties of hits and leads. Development of the methods reviewed in this paper concerning log P prediction seems to be a significant tendency in the modern pharmaceutical industry.

European Journal of Medicinal Chemistry, 2010

Simplified molecular input line entry system (SMILES) has been utilized in constructing quantitative structureeproperty relationships (QSPR) for octanol/water partition coefficient of vitamins and organic compounds of different classes by optimal descriptors. Statistical characteristics of the best model (vitamins) are the following: n ¼ 17, R 2 ¼ 0.9841, s ¼ 0.634, F ¼ 931 (training set); n ¼ 7, R 2 ¼ 0.9928, s ¼ 0.773, F ¼ 690 (test set). Using this approach for modeling octanol/water partition coefficient for a set of organic compounds gives a model that is statistically characterized by n ¼ 69, R 2 ¼ 0.9872, s ¼ 0.156, F ¼ 5184 (training set) and n ¼ 70, R 2 ¼ 0.9841, s ¼ 0.179, F ¼ 4195 (test set).