Research Articles
Prediction of the aqueous solubility: Comparison of the general solubility equation and the method using an amended solvation energy relationship

https://doi.org/10.1002/jps.10022Get rights and content

Abstract

An Amended Solvation Energy Relationship (ASER) was recently reported to successfully predict the aqueous solubilities of a set of 664 organic compounds. The average absolute error and root mean square error are 0.43 and 0.62 log units, respectively. When the General Solubility Equation (GSE) is applied to the same set of compounds, it gives an average absolute error of 0.45 log units and a root mean square error of 0.62 log units. These results are similar to those of the ASER method. The advantages and disadvantages of each method are discussed. It is shown that when the two methods agree with each other, they also agree with the experimentally determined values.

Section snippets

INTRODUCTION

Aqueous solubility is a crucial physical property in pharmaceutical and environmental research. Several methods for the prediction of the aqueous solubility have been published in recent years.1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11., 12., 13. The General Solubility Equation (GSE), as initially proposed by Yalkowsky and Valvani in 1980,2 and recently revised by Jain and Yalkowsky,13 has been used widely. The GSE relates the molar aqueous solubility (Sw) to the Celsius melting point (mp) and

METHODS

The melting point data for the 664 compounds were obtained from the AQUASOL dATAbASE, Merck Index, and several Internet databases. The octanol–water partition coefficients were calculated with CLOGP® software (Version 4.0, BioByte Corp., Claremont, CA). Experimental octanol–water partition coefficients are listed if available. The experimental aqueous solubilities and the ASER predicted solubilities are those reported by Abraham and Le.14 The aqueous solubilities were also calculated using the

Octanol–Water Partition Coefficient

Experimental partition coefficients were found for 530 of the 664 compounds. As shown in Figure 1, the calculated partition coefficients using CLOGP® are in very good agreement with the available measured values (MLOGP), with an AAE of only 0.121 log units. CLOGP® version 4.0 seems to give more accurate estimations of octanol–water partition coefficients than the previous version. Partition coefficients calculated with CLOGP® are used in the solubility calculations because they are easily

DISCUSSION

In spite of the fact that the GSE uses only two input variables and does not use any training set or fitted parameters, it gives quite reasonable predictions. The ASER method produces only marginally better predictions (Table 1). However, the ASER method uses seven coefficients, six variables derived from a very large number of structural descriptor values, and multiple linear regression analysis.

As can be seen in Table 1, both methods give better prediction for the 408 liquids than for the 256

CONCLUSIONS

Both the ASER and the GSE methods give satisfactory prediction for the compound set. The ASER is based on multiple linear regression analysis of a large training set that may or may not contain the required structural fragments, whereas the GSE is simpler and more user‐friendly. However, the latter requires knowledge of either an experimentally determined or an estimated mp of the solute. This study provides support for the reliability of the GSE in estimating the aqueous solubilities of

REFERENCES (21)

There are more references available in the full text version of this article.

Cited by (52)

  • Determining the solubilities for benzoate, nicotinate, hydrochloride, and malonate salts of bedaquiline

    2022, International Journal of Pharmaceutics
    Citation Excerpt :

    The aqueous solubility values for the salts predicted with the GSE were much lower than the actual lab-generated data. This variability was consistent with the experimental error values of between 0.4 and 0.6 log molar solubility units for some pharmaceutical compounds reported by Yalkowsky et al (Ran, Jain, & Yalkowsky, 2001; Yang, Ran, & Yalkowsky, 2002). However, aqueous solubility may only be important for drugs that will be formulated as injectables (Byrn, Zografi, & Chen, 2017), and may not play a major role in the lipid bi-layer of the human GIT.

  • Encapsulation, protection, and release of hydrophilic active components: Potential and limitations of colloidal delivery systems

    2015, Advances in Colloid and Interface Science
    Citation Excerpt :

    Saturation concentrations can sometimes be estimated from knowledge of the molecular properties of the solutes and solvents involved using empirical, theoretical, or computational models [37–39]. For example, the General Solubility Equation (GSE) gives a reasonable estimate of the water-solubility of small solutes based on knowledge of a few of their physicochemical properties: log(CW) = 0.5 − 0.01 (Tm − 25) − log(KOW), where CW is the molar water-solubility, Tm is the melting point of the solute, and KOW is the oil–water partition coefficient of the solute [40,41]. The solubility and partitioning of many hydrophilic molecules are highly dependent on pH because they have dissociable ionic groups, such as carboxyl (–COOH ↔ –COO− + H+) and amino groups (–NH3+ ↔ NH2 + H+).

  • Crystals and crystallization in oil-in-water emulsions: Implications for emulsion-based delivery systems

    2012, Advances in Colloid and Interface Science
    Citation Excerpt :

    At present no single approach can accurately predict the solubility of all solutes in all solvents, but many of the methods do produce fairly good estimations for small solutes. An example of a relatively simple method that can be used to relate the water-solubility of solutes to easily measurable properties is the General Solubility Equation (GSE): log(SW) = 0.5 − 0.01 (Tm − 25) − log(KOW), where SW is the molar water-solubility, Tm is the melting point, and KOW is the oil–water partition coefficient of the solute [33,34]. This model indicates that the water-solubility of a substance should decrease as its melting point and KOW increase.

View all citing articles on Scopus
View full text