# Weighing up ways of working out log *P*

## Ezine

**Published:**Jul 15, 2017**Author:**Ryan De Vooght-Johnson**Channels:**

## Numerous methods of measuring and calculating lipophilicity

Lipophilicity, or the ability of a molecule to dissolve in non-polar (‘fatty’) solvents, is an important variable affecting drug compounds’ activities. Lipophilicity is typically measured by log *P* values, which are the logarithms of the partition coefficient between n-octanol and water.

Direct measurement of log *P* using shaking flasks is possible, but suffers from problems such as the limited range of possible concentrations, possible emulsion formation, the need for significant amounts of compound,. In practice other methods, such as reversed-phase high-performance TLC (RP-HPTLC), reversed-phase HPLC and microemulsion liquid chromatography (MELC), are often used.

RP-HPTLC gives rise to a number of lipophilicity parameters that are calculated from the Rf values in a particular solvent. These include the Rm value (retention parameter, given by the equation Rm = log(1/R_{f} – 1)); the Rm0 value (the Rm value extrapolated for pure water); the b value (the slope of the equation Rm = Rm^{0} + bφ, where φ is the volume fraction of the organic solvent in the mobile phase) and the C_{0} value (given by the equation C_{0} = - Rm^{0} / b), which can be considered to be the hydrophobicity per unit of hydrophobic surface.

In addition to experimental measurements, there are a host of computational programmes that give lipophilicity values. The University of Belgrade researchers aimed to apply several different methods, both experimental and computational, to determine the lipophilicity of a series of 17 benzothiepino[3,2-c]pyridine derivatives, referred to as compounds 1–17, being investigated as potential antifungal drugs.

## Lipophilicity determined by a combination of different methods

Reversed-phase high-performance TLC was carried out using Merck RP-18W F254s plates in a horizontal developing chamber. Three mixtures, aqueous methanol, aqueous acetone and aqueous dioxane, were used, each at a number of different organic solvent to water ratios. The mean of the Rm values for each solvent, mRm, was calculated. For MELC, two different solvents systems were used, both consisting of mixtures of n-heptane, 1-butanol, water and SDS (sodium dodecyl sulphate).

Eight different theoretical log *P* values, along with one water solubility value, were calculated for each compound using various software programmes. Subsequent statistical calculations were carried out using Statistica 10 (StatSoft) and Microsoft Excel.

Principal component analysis (PCA) was carried out on the tabulated data from all methods for all 17 compounds, giving one main component accounting for 81.2% of the variability and a second component accounting for 7.4% of the variability. The PCA plot showed no clear ‘most lipophilic’ compound, with a number of compounds showing relatively high lipophilicity. The PCA loading plot showed that the variables b and Rm_{0} from the methanol TLC gave values closest to the arithmetic mean.

Sum of Ranking Differences (SRD) scores were calculated; these are a measure of how much the values from a particular method differ from a reference value, which in this case was the arithmetic mean from all the methods. A comparison of SRD scores showed that Xlog*P*2 (a computational method) and PC1 from the acetone TLC values were closest to the overall arithmetic mean. PC1 from dioxane TLC values was shown to be the least accurate method, ranking the compounds no better than chance.

The most lipophilic compound was calculated as compound 4 from the chromatographic data, but as compound 10 from the computational data. Statistical examination of SRD values confirmed compound 4 as the most lipophilic. It was also the compound that showed the greatest antifungal activity against *C. Albicans*, a type of yeast from the human gut that can cause disease in immunocompromised people.

## 'Consensus' log *P* values obtained by statistical methods

The authors present a comprehensive statistical treatment of the log *P* values determined by different methods. The use of differences from the overall mean enable a ‘consensus’ decision on the correct order of lipophilicity to be reached. The method could doubtless be extended to other series of active compounds.

* *

* *

**Follow us on Twitter!**