Hi, Thanks for the reply. I've looked into the code and although it partially solves the problem, it does so by forcing the user to split the SMARTS query into each atom and manually bond them. I'm trying to use some SMARTS definitions from a commercial package directly without modifications rendering this strategy unusable.
If I may suggest and improvement to the CDK it would definitely be to respect the ordering of atoms returned by a SMARTS query. Right now the SMARTS matcher can be used for identifying how many features of a type there is in a molecule and where they are. It is not possible to extract information about what the functional parts of the features are as the ordering of the returned atoms is arbitrary. From having looked into the code I'd say that changing this is a major restructuring of how the code currently works, but I could be wrong. Once again, thanks for building and maintaining the CDK. Thomas On Wed, Sep 15, 2010 at 12:02 AM, Rajarshi Guha <[email protected]> wrote: > I don't have access to the sources right now, so I don't exactly remember > how i did it, but the CDK pharmacophore matching code is doing this (ie > identifying atoms matching the pharmacophore SMARTS pattern). > > On Sep 14, 2010, at 5:12 AM, Thomas G. Kristensen wrote: > >> Hi all, >> >> I'm trying to use CDK for detecting pharmacophores. It's easy enough >> to get SMARTSQuery to find patterns but a major problem is, that the >> returned atom (indices) do not respect the ordering of molecules in >> the original SMARTS string. For example, the query string "[O;X2]C" >> will, in one of my examples, return two atoms in this order: >> >> #<Atom Atom(745381933, S:C, 3D:[(-4.8833, 0.8605, 0.0308)], >> AtomType(745381933, N:C.sp3, FC:0, H:SP3, NC:4, EV:4, >> Isotope(745381933, Element(745381933, S:C, ID:C1))))> >> #<Atom Atom(1362034980, S:O, 3D:[(-3.6751, 1.6658, 0.029)], >> AtomType(1362034980, N:O.sp3, FC:0, H:SP3, NC:2, EV:2, >> Isotope(1362034980, Element(1362034980, S:O, ID:O1))))> >> >> That is, the carbon is returned before the oxygen. I'm only interested >> in the oxygen as that is the defining point of this feature but I >> cannot assume that it will be the first atom returned (or the last, >> for that matter). The order in which SMARTSQuery returns matched atoms >> is arbitrary, rendering it useless as a feature extraction tool. >> >> I have been looking trough the source code of SMARTQuery, SMARTSParser >> and UniversalIsomorphismTester to no avail. As far as I can tell from >> the source it is impossible to extract the atoms in the correct order >> without completely rewriting the SMARTS matching functionality include >> the isomorphism tester. >> >> Does anybody know if I am mistaken? It would really make life much >> easier for me if it was possible to extract the atoms in the order in >> which they are matched. >> >> Best regards, >> >> Thomas >> >> >> ------------------------------------------------------------------------------ >> Start uncovering the many advantages of virtual appliances >> and start using them to simplify application deployment and >> accelerate your shift to cloud computing. >> http://p.sf.net/sfu/novell-sfdev2dev >> _______________________________________________ >> Cdk-user mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/cdk-user > > ---------------------------------------------------- > Rajarshi Guha | NIH Chemical Genomics Center > http://www.rguha.net | http://ncgc.nih.gov > ---------------------------------------------------- > After a number of decimal places, nobody gives a damn. > > > ------------------------------------------------------------------------------ Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev _______________________________________________ Cdk-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cdk-user
