Prediction of Partition Coefficients of Environmental Toxins Using Computational Chemistry Methods

Abstract

The partitioning of compounds between aqueous and other phases is important for predicting toxicity. Although thousands of octanol–water partition coefficients have been measured, these represent only a small fraction of the anthropogenic compounds present in the environment. The octanol phase is often taken to be a mimic of the inner parts of phospholipid membranes. However, the core of such membranes is typically more hydrophobic than octanol, and other partition coefficients with other compounds may give complementary information. Although a number of (cheap) empirical methods exist to compute octanol–water (log k OW ) and hexadecane–water (log k HW ) partition coefficients, it would be interesting to know whether physics-based models can predict these crucial values more accurately. Here, we have computed log k OW and log k HW for 133 compounds from seven different pollutant categories as well as a control group using the solvation model based on electronic density (SMD) protocol based on Hartree–Fock (HF) or density functional theory (DFT) and the COSMO-RS method. For comparison, XlogP3 (log k OW ) values were retrieved from the PubChem database, and KowWin log k OW values were determined as well. For 24 of these compounds, log k OW was computed using potential of mean force (PMF) calculations based on classical molecular dynamics simulations. A comparison of the accuracy of the methods shows that COSMO-RS, KowWin, and XlogP3 all have a root-mean-square deviation (rmsd) from the experimental data of ≈0.4 log units, whereas the SMD protocol has an rmsd of 1.0 log units using HF and 0.9 using DFT. PMF calculations yield the poorest accuracy (rmsd = 1.1 log units). Thirty-six out of 133 calculations are for compounds without known log k OW , and for these, we provide what we consider a robust prediction, in the sense that there are few outliers, by averaging over the methods. The results supplied may be instrumental when developing new methods in computational ecotoxicity. The log k HW values are found to be strongly correlated to log k OW for most compounds.

Local Anesthetics Transfer Across the Membrane: Reproducing Octanol-Water Partition Coefficients by Solvent Reaction Field Methods

“…When determining the properties of other small organic compounds, the SMD model has proven to be the most precise solvation method as well. [34][35][36] In this respect, we recommend application of the SMD solvation model during the process of designing novel local anesthetics. The COSMO-RS solvation model, which also proved to yield comparably reliable n-octanol/water partition coefficients, is unfortunately not implemented in Gaussian 16.…”

“…The COSMO-RS solvation model, which also proved to yield comparably reliable n-octanol/water partition coefficients, is unfortunately not implemented in Gaussian 16. 35 It should be emphasized that among the applied solvent reaction field methods only SMD was parametrized to reproduce solvation free energy in n-octanol. The critical elements of parametrization are always nonbonding interactions with the solvent that include reversible work for cavity creation, a dispersion component of the solvation energy and an interaction between the charge distribution and dielectric continuum.…”

Predicting Octanol–Water Partition Coefficients of Fluorinated Drug-Like Molecules: A Combined Experimental and Theoretical Study

“…As shown, the best performing empirical methods are miLOGP, ALOGPS, and ALOGP where the deviations are consistently below 0.5 log units. Interestingly, XLOGP3, which was previously identified as a reasonably robust method for predicting the log P of simple fluorinated compounds, [20][21][22] displayed a mean absolute deviation (MAD) of 0.82 and maximum absolute deviation (AD max ) of 1.25 log units, which is beyond the acceptable error range. This highlights one of the limitations of empirical methods as their performance can be less predictable when they are applied to molecules dissimilar to those in their training set, e.g.…”

“…The results here echo the observations from several recent studies where physics-based implicit and explicit solvent models do not necessarily yield more accurate log P predictions compared with empirical models. [20,21] An important consideration when evaluating the performance of empirical models is whether there is significant overlap between the test and training sets. Given that this is the first time experimental log P values have been reported for the novel fluorinated molecules in this study, the miLOGP, ALOGP, and ALOGPS can be considered to be reasonably robust.…”

“…[19] All three classes of methods, particularly fragment-based methods, have been widely used for the prediction of log P, although direct comparisons of these methods on a common dataset is relatively scarce. Recent work by the Ho and Spoel groups have indicated that empirical methods perform unexpectedly well, even when applied to relatively 'exotic' molecules outside of their training set, [20,21] and that their performance are comparable if not superior to physics-based implicit and explicit solvent models. Similar observations were made in the very recent SAMPL6 Partition Coefficient Challenge where several empirical methods ranked highly in this challenge.…”