Drug-logS (prediction model of solubility) is a solubility prediction program solely based on simple atomic contribution. This version is different from the old one reported in reference [1] (the old version was implemented in MOE). The atom types used in the old version were re-optimized using the database reported in reference [2]. The number of atom types used in the new model is 64, rather than 76 in the old model. Using the new model, the prediction of the 1290 molecules used in reference [1] can be improved obviously, as indicated by a standard deviation (s) of 0.39 and an unsigned mean error (UME) of 0.46. In the new version, two correction factors (Hydrophobic Carbon and square of molecular weight) were also employed. It should be noted that the definition of Hydrophobic Carbon is a little different (see user guide).

Note:

The above program is password protected when compressed by using zip. Sign the license file and apply for the passwd from the ADME group.


Download:

License file: HlogS_license.pdf | HlogS_license.doc

Source code and manual: HlogS.tar.zip | HlogS_manual.pdf


Reference:

[1]. Tingjun Hou, Ke Xia, Wei Zhang, Xiaojie Xu, ADME evaluation in drug discovery. 4. Prediction of aqueous solubility based on atom contribution approach, Journal of Chemical Information and Computer Sciences, 2004, 44, 266-275.

[2]. Junmei Wang, George Krudy, Tingjun Hou, George Holland, Xiaojie Xu, Development of reliable aqueous solubility models and their application in drug-like analysis, Journal of Chemical Information and Modeling, 2007, 47, 1395-1404.