Analysis of water solubility data on the basis of HYBOT descriptors

Abstract

Solubility data of 787 organic liquids (electrolytes and nonelectrolytes) with diverse structures has been quantitatively described by physicochemical property descriptors. Special effects like intra-and intermolecular hydrogen bonds have been shown to be very important for water solubility. It is found that an important part of the solutesolvent interaction is neglected in all correlations of logS with (only) logP, as in this case the solute H-bond donor effect is not considered. As expected intramolecular hydrogen bonds lead to reduced solubility, whereas intermolecular hydrogen bonds (both HB donors and acceptors) of solutes result in higher solubility. An exception to the latter rule are carboxylic acids which due to intermolecular HB-induced dimerization in the pure liquid phase of acids show a three times lower solubility as expected on the basis of their molecular properties. A volume-related term (molecular polarizability a) was found to have an essential negative contribution to solubility. For the first time the solubility increasing effect of partial ionization of weak acids and bases in saturated aqueous solutions has been quantitatively considered for sets of compounds by exact calculation of the pH determined by the solutes aqueous solubility and pKa value(s).

Recent progress in the computational prediction of aqueous solubility and absorption

“…Two very interesting recent papers applied the HYBOT descriptors to solubility prediction of liquids 37 and solids. 36 They are both discussed in greater detail below, in the section about crystal effects on solubility prediction.…”

“…37 A relatively small portion of their training data were ionizable compounds with no reported pH information. Their assumption was that the solubility measurements were done in unbuffered solution.…”

“…36 , 37 Using a collection of liquid chemicals and drugs, 37 they demonstrate that descriptors previously identifi ed [67][68][69][70] as being relevant for logP prediction perform admirably in the GSE in place of the logP term. This is an interesting result, as hydrogen bond donor strength was not an important descriptor for logP prediction but appears critical for modeling solubility.…”

Acute intravenous toxicity to mice calculations on the basis local regression models in superoverlapping clusters (LRMSC)

“…Указанный подход был успешно применен для расчета липофильности [2], растворимости в воде [3,4], адсорбции химических соединений в организме человека [5,6], и токсичности по отношению к Guppy [7]. Опыт использования уравнения (2) показывает, что результаты расчета становятся лучше в случае, когда структурный сосед имеет близкие физико-химические свойства по отношению к рассматриваемому соединению.…”

QSPR Prediction of Aqueous Solubility of Drug-Like Organic Compounds

“…[20][21][22] Several articles have published with MLR models for the prediction of aqueous solubility. [23][24][25][26] In a QSPR study, a mathematical model is developed which relates the structure of a set of compounds to a physical property such as aqueous solubility. In a QSPR study is that there is some sort of relationship between the physical property of interest and structural descriptors.…”