IMPROVED METHOD FOR ESTIMATING WATER SOLUBILITY FROM OCTANOL WATER PARTITION-COEFFICIENT

Water solubility is a critical property in risk assessments for chemic als, but measured values are often unavailable. The most practical app roach to estimating solubility for diverse and complex structures invo lves regression against the logarithm of the octanol/water partition c oefficient (log K-ow). However, currently used equations were develope d from relatively small training sets (200-300 compounds). Our objecti ve was to upgrade log K-ow-based methods by examining a much larger an d diverse training set and, if possible, to improve estimation accurac y further by incorporating a series of simple correction factors. We a lso evaluated the contributions of melting temperature (T-m) and molec ular weight (MW). Measured values of water solubility, log K-ow and T- m (solids only) were available for all compounds in the training set ( n = 1,450). The best equation for this dataset included log K-ow, T-m, MW, and 12 correction factors as independent variables, with r(2) = 0 .970, SD = 0.409, ME = 0.313. Statistics for an independent validation set of 817 compounds were consistent with these results: r(2) = 0.902 , SD = 0.615, ME = 0.480. We found that T-m contributed significantly to estimation accuracy for solids; addition of MW also increased accur acy; measured values of log K-ow, and T-m produced more accurate solub ility estimates than did estimated values; and the new method outperfo rmed other widely used general-purpose equations based on log K-ow.