Solubility Prediction Methods

An extensive series of studies for the prediction of aqueous solubility has been reported in the literature, as summarized by Lipinski et al. [15] and jorgensen and Duffy [16]. These methods can be categorized into three types 1 correlation of solubility with experimentally determined physicochemical properties such as melting point and molecular volume 2) estimation of solubility by group contribution methods and 3) correlation of solubility with descriptors derived from the molecular structure by computational methods. The third approach has been proven to be particularly successful for the prediction of solubility because it does not need experimental descriptors and can therefore be applied to collections of virtual compounds also. 

1 Correlation with Descriptors from Experimental Data 

A series of studies has been made by Yalkowsky and co-workers. The so-called general solubility equation was used for estimating the solubility of solid nonelectrolytes [17, 18]. The solubility log S (logarithm of solubility expressed as mol/L) was formulated with log P logarithm of octanol/water partition coefficient), and the melting point (MP) as shown in Eq. (11). This equation generally 

It is remarkable that only two descriptors were needed in this method. However, this equation is mostly only of historical interest as it is of little use in modem dmg and combinatorial library design because it requires a knowledge of the compound s experimental melting point which is not available for virtual compounds. Several methods exist for estimating log P [1-14], but only a few inroads have been made into the estimation of melting points. The melting point is a key index of the cohesive interactions in the solid and is still difficult to estimate. 

The group contribution method allows the approximate calculation of solubility by summing up fragmental values associated with substmctural units of the compounds (see Section 7.1). In a group contribution model, the aqueous solubility values are computed by Eq. (12), where log S is the logarithm of solubility, C is the number of occurrences of a substmctural group, i, in a molecule, and is the relative contribution of the fragment i. 

---

It the productivity target cannot be achieved then a co-solvent system could be selected using solubility prediction methods like NRTL-SAC [1] and Local UNIFAC [4], The addition of a second solvent to increase solubility is an effective way of increasing productivity for a sparingly soluble solute. 

In special cases, polar gases such as ammonia, formic acid and water are doped into the carrier gas to improve the analyte s solubility in the carrier gas. In both supercritical fluid and liquid chromatography, the analyte solubility in the carrier liquid affects the retention time. The carrier liquid is called the eluent and/or the mobile phase. The prediction of retention times in liquid chromatography is very difficult due to the lack of a solubility prediction method. However, the retention can be predicted by computational chemical methods using model phases. ... 

QSAJi Methods for Fluid Solubility Prediction. Several group contribution methods for predicting Hquid solubiHties have been developed. These methods as weU as other similar methods are often called quantitative stmcture-activity relationships (QSARs). This field is experiencing rapid development. 

Ktihne, R., Ebert, R-U., Schtitirmann, G. Model selection based on structural similarity - method description and application to water solubility prediction. f Chem. Inf Model. 2006, 46, 636-641. 

In this book we will focus on physicochemical profiling in support of improved prediction methods for absorption, the A in ADME. Metabolism and other components of ADME will be beyond the scope of this book. Furthermore, we will focus on properties related to passive absorption, and not directly consider active transport mechanisms. The most important physicochemical parameters associated with passive absorption are acid-base character (which determines the charge state of a molecule in a solution of a particular pH), lipophilicity (which determines distribution of a molecule between the aqueous and the lipid environments), solubility (which limits the concentration that a dosage form of a molecule can present to the solution and the rate at which the molecule dissolves from... 

As a key first step towards oral absorption, considerable effort has been directed towards the development of computational solubility prediction [26-30]. However, partly due to a lack of large experimental datasets measured under identical conditions, today s methods are not sufficiently robust for reliable predictions [31]. Nonetheless, further fine-tuning of these models can be expected since high-throughput data have become available for their construction. 

The octanol-water partition coefficient Kow is widely used as a descriptor of hydrophobicity. Variation in /fow is primarily attributable to variation in activity coefficient in the aqueous phase (Miller et al. 1985) thus, the same correlations used for solubility in water are applicable to /fow. Most widely used is the Hansch-Leo compilation of data (Leo et al. 1971, Hansch and Leo 1979) and related predictive methods. Examples of Kow correlations are ... 

It is difficult to accurately predict aqueous solubility from chemical structure, because it involves disruption of the crystal lattice as well as solvation of the compound. Simple methods based on log P and melting temperature have been widely used [113, 114]. Recently, various prediction methods have been reported [115-125] that are able to predict aqueous solubility to within ca. 0.5 log units (roughly a factor of 3 in concentration). Although these predictors may not be precise or robust enough to select final compounds, they can be used as rough filters for narrowing the list of candidates. 

We have studied the performance of several prediction methods to see how well in-house thermodynamic solubility measurements could be predicted. Among the prediction methods we studied were Huuskonen s method [26], ACD/Solubility DB [38], Meylan s method [21] as implemented in QMPRPlus, and the SimulationsPlus solubility prediction as implemented in QMPRPlus [39]. In general, we foimd the predictions to... 

There are several properties of a chemical that are related to exposure potential or overall reactivity for which structure-based predictive models are available. The relevant properties discussed here are bioaccumulation, oral, dermal, and inhalation bioavailability and reactivity. These prediction methods are based on a combination of in vitro assays and quantitative structure-activity relationships (QSARs) [3]. QSARs are simple, usually linear, mathematical models that use chemical structure descriptors to predict first-order physicochemical properties, such as water solubility. Other, similar models can then be constructed that use the first-order physicochemical properties to predict more complex properties, including those of interest here. Chemical descriptors are properties that can be calculated directly from a chemical structure graph and can include abstract quantities, such as connectivity indices, or more intuitive properties, such as dipole moment or total surface area. QSAR models are parameterized using training data from sets of chemicals for which both structure and chemical properties are known, and are validated against other (independent) sets of chemicals. 

Li, A., and A. W. Andren, Solubility of polychlorinated biphenyls in water/alcohol mixtures. 2. Predictive methods , Envion. Sci. Technol., 29, 3001-3006 (1995). 

Substances such as PCBs can have activity coefficients exceeding 1 million. Hydropho-bicity thus is essentially an indication of the magnitude of y. Some predictive methods focus on estimating y, from which solubility can be deduced. 

Rose, F L., Mclntire, C.D. (1970) Accumulation of dieldrin by benthic algae in laboratory streams. Hydrobiologia 35, 481. Rothman, A.M. (1980) Low vapor pressure determination by the radiotracer transpiration method. 7. Agric. Food Chem. 28,1225-1228. Ruelle, R, Kesselring, U.W. (1997) Aqueous solubility prediction of environmentally important chemicals from the mobile order thermodynamics. Chemosphere 34(2), 275-298. 

---