aqueous solubilities of organic compounds



Dear collegues,
 as I promised, I provide you with a summary of replies, that I received upon
 my request about methods for the prediction of aqueous solubilities.
 I would like to thank everyone, who responded !
 Best regards,
 Christian
 ________________________________________________________________
    Dr. Christian Pilger
    Dept. Chemical Research / Structural Research
    K91-00-10
    Boehringer Ingelheim Pharma KG
    D-88397 Biberach/Germany
    Phone: 07351-545749
    Fax:     07351-5497924
    mailto: christian.pilger ( ( at ) ) bc.boehringer-ingelheim.com
 ---------------------------
 Carlos Faerman [CFaerman ( ( at ) ) etrx.com]
 Dear Dr. Pilger,
 This is a very important issue in the pharma/biotech world !
 I have two remarks to make:
 1) The experimental values for solubilities are available, among other
 sources, in the Beilstein database, currently available online.
 2)Without endorsing any specific software company, Gastro Plus by
 Simulations Plus, Lancaster California seems to have a reasonable
 success in predicting solubility.
 I would be very interested in finding out what your experience or
 someone else's is in this field.
 ---------------------------
 David Gallagher [dgallagher ( ( at ) ) cachesoftware.com]
 There is a zipped file which includes a water solubility QSPR  along with
 the calibration set of 156 environmentally significant organic molecules
 and water solubility data available at:
 http://www.cachesoftware.com/downloads/LogW.zip The QSPR
 uses descriptors
 calculated by MOPAC-COSMO. The r^2 correlation is 0.95.
 It is located on: http://www.cachesoftware.com/cache/applications.shtml If
 you do not already have the CAChe software to open the files, you can
 download a free evaluation copy from the same site.
 David Gallagher, Fujitsu
 ---------------------------
 David van der Spoel [spoel ( ( at ) ) xray.bmc.uu.se]
 Check out:
  ( ( at ) ) Article{Wang2001b,
   author = 	 {J. M. Wang and W. Wang and S. H. Huo and M. Lee and P. A.
 Kollman},
   title = 	 {Solvation model based on weighted solvent accessible
 surface area},
   journal = 	 {J. Phys. Chem. B},
   year = 	 2001,
   volume =	 105,
   pages =	 {5055-5067}
 }
 A very simple area based method. However, I haven't been able to get
 meaningful numbers out of it..
 ---------------------------
 Dearden John [J.C.Dearden ( ( at ) ) livjm.ac.uk]
 We have recently tested the ability of a number of software packages to
 predict aqueous solubility, and I attach a few PowerPoint slides with some
 preliminary results. We took our 113 compounds from a data-set of over 600
 solubilities published by Abraham and Le (J. Pharm. Sci. 88 (1999) 868-880,
 by taking roughly every sixth compound. I've included a similar comparison
 of log P predictions, from which you can see that the ability to predict log
 S is nothing like so good as the ability to predict log P.
 ---------------------------
 Erik Johansson [erik.johansson ( ( at ) ) umetrics.com]
 Arne Brändström (retired from AZ Mölndal) has worked a lot on his
 own method
 for aqueous solubility.
 I saw a early manuscript a couple of years ago but I am not sure if this is
 published.
 ---------------------------
 George Vacek [vacek ( ( at ) ) schrodinger.com]
 QikProp predicts a variety of pharmaceutical properties, including
 solubility,
 that are
 important in determining ADME characteristics for drug candidates.  The
 complete
 set
 of  properties predicted by QikProp, methods, data and literature references
 are
 available at
 http://www.schrodinger.com
 For the specific case of solubility, the data set and descriptors are
 detailed in
 the article
 W. L. Jorgensen and E. M. Duffy, Bioorg. Med. Chem. Lett., 10 (2000) 1155-8.
 ---------------------------
 Igor Tetko [itetko ( ( at ) ) yahoo.com]
 You can use our ALOGPS software that predict both aqueous  solubility
 and lipophilicity of chemicals. The description of this software can
 be found at:
 Tetko, I.V.; Tanchuk, V. Yu.; Kasheva, T. N.; Villa, A. E. P.
 Estimation of Aqueous Solubility of Chemical Compounds Using E-state
 Indices, J. Chem. Inf. Comput. Sci., 2001, 41, 1488-1493.
 Tetko, I.V.; Tanchuk, V. Yu.; Villa, A. E. P. Prediction of
 n-Octanol/Water Partition Coefficients from PHYSPROP Database Using
 Artificial Neural Networks and E-state Indice, J. Chem. Inf. Comput.
 Sci., 2001, 41, 1407-1421.
 Tetko, I.V.; Tanchuk, V.Yu.; Kasheva, T.N.; Villa, A.E.P. Internet
 Software for Calculation of Lipophilicity and Aqueous Solubility of
 Chemical Compounds, J. Chem. Inf. Comput. Sci., 2001, 41, 246-252.
 We have new version 2.1 of this program that is available for testing
 as a standalone program. You can also use free on-line version ALOGPS
 2.0 available at WWW http://www.lnh.unil.ch/~itetko/logp. A nice
 feature of the standalone version is a possibility to create a user's
 LIBRARY  of compounds. This can improve prediction of this software
 for similar compounds in 2-3 times (article in preparation). The
 ALOGPS 2.1 will be available at the WWW soon. The description of the
 algorithm used in the LIBRARY mode, Associative Neural Network, will
 be published in JCICS soon (I am waiting for the galley-proof of the
 article).
 You can download the set of compounds used in the aqueous  solubility
 study at  http://www.lnh.unil.ch/~itetko/logp/logs.html
 ---------------------------
 John McKelvey [jmmckel ( ( at ) ) attglobal.net]
 I think that one of the best beople to talk to about this might be Andreas
 Klampt there in Germany...Andreas.Klamt ( ( at ) ) CosmoLogic.de
 ---------------------------
 Hugo Kubinyi  [kubinyi ( ( at ) ) t-online.de]
 Dear Dr. Pilger,
 there is a simple solution for liquids - just take the inverse lipophilicity
 as a good estimate. The problem is much more difficult for solids. Most
 probably you know about the correlation of solubility with log P and mp. -
 however there is no method to predict the mp., the crystal lattice, the
 intermolecular forces in the crystal and, therefore, no possibility to
 predict the solvation enthalpy and entropy - which influences lipophilicity.
 If somebody (especially vendors) tells you something else, you should be a
 little suspicious.
 For a database that is not too expensive see:
 http://esc.syrres.com/interkow/PhysProp.htm
 ---------------------------
 Robert L. Lipnick [Lipnick.Robert ( ( at ) ) epamail.epa.gov]
 Check papers on water solubility estimation from chemical structure and log
 P and MP by Samuel Yalkowsky, Univ. of Arizona. Also, he maintains a water
 solubility database.
 ---------------------------
 Lowell Hall [hall ( ( at ) ) enc.edu]
 Dear Christian:
 I collaborate with Marc Parham of Interactive analysis, Bedford, MA.  We
 have a model that is available for evaluation on a web site.  The model is
 based on a very large database of experimental data.  You can contact Marc
 at
 Marc Parham <mparham ( ( at ) ) gis.net>
 and view the site at
 www.InteractiveAnalysis.com
 The computation is very fast.
 ---------------------------
 Phil Howard [howardp ( ( at ) ) syrres.com]
 We have a program that predicts water solubility from
 octanol/water partition coefficient (see
 http://esc.syrres.com/interkow/wskow.htm) and it is
 available for free from
 the U.S.EPA (http://www.epa.gov/oppt/exposure/docs/episuitedl.htm). We
 are
 working on a program that will calculate water solubility using fragment
 constants that will be available soon.  One of the nice features of this
 program is that it will allow one to calculate the water solubility from a
 close structural analog by using only the fragments that are different
 between the structural analog and the chemical of interest.
          We also have a database of water solubility data (see
 http://esc.syrres.com/interkow/PhysProp.htm) that has over
 6000 water
 solubility values.  This is available in a variety of formats including
 ISISBase; the latter allows substructure searching.
 ---------------------------
 J. W. McFarland [reckon.dat ( ( at ) ) attglobal.net]
 Dear Dr. Pilger:
 Here is a recent publication that describes one method and refers to some
 others
 that have been published. Some of these contain extensive datasets. Others
 have
 been published since. If you send me your address I will send you a reprint.
 McFarland, J. W.; Avdeef, A.; Berger, C. M.; Raevsky, O. A. Estimating the
 Water
 Solubilities of Crystalline Compounds from Their Chemical Structures Alone.
 J.
 Chem. Inf. Comput. Sci. 2001, 41(5), 1355-1359.
 ---------------------------
 Stephen  Bowlus [stephen.bowlus ( ( at ) ) lionbioscience.com]
 The most extensive commercial effort I am aware of is ACD, which uses
 variants of the general solubility equation, with terms as surrogates for
 melting point where this is not known.
 Most of the models of which I am aware are based on the Yalkowsky model, or
 that is the one most frequently discussed in the literature.  The main
 problem with it seems to be the melting point:  there is limited utility to
 the model if you can't get an accurate estimate (i.e. measurement) of mp.
 So within this framework, methods to estimate mp are of great interest.
 A frequently overlooked program in this area is a group contribution method,
 Chemicalc (Suzuki & Kudo, JCAMD, 1990, 4, 155; Suzuki, JCAMD, 1991, 5, 149).
 I have used this mostly for logP estimation, so I can't comment on its
 strength for solubility, which it also calculates.  The program is available
 > from QCPE, or I can send you the code for the PC (runs in a DOS window).