On 17 December 2011 04:46, Geoff Hutchison <ge...@geoffhutchison.net> wrote: >> The weird thing is that IdentifyResidue can identify the first serine >> without any problems; it just doesn't identify the second one. Also, >> it will identify them both correctly if the --gen3d is omitted (or >> more precisely, if the call to Builder.Build is omitted). > > This bug report exposed two issues: > 1) Chain / residue perception is brittle with respect to bond indexes (e.g., > if the beta carbon has a bond to the side chain first or second) > 2) OBBuilder::Build re-orders bonds (i.e., it deletes them all, creates some, > and then adds the rest at the end) > > The most immediate fix should be to address #2 -- if the bonds are kept in > the same order, the bug should disappear. > > The problem is that it leaves chains.cpp in a very brittle state -- > reordering atoms or bonds would still cause hard-to-debug failures in residue > perception. (Come to think of it, this explains several weird bug reports.) > > I will investigate a more permanent solution, but it may take time. This is > opaque OpenEye code with poor commenting. The key problem seems to be in how > the TracePeptideBackbone() function works -- it probably needs to be more > flexible, particularly with C(O) atoms, which could either be SER side-chains > or backbone atoms.
That's great Geoff. And I think I saw that you commited a fix to Builder. - Noel ------------------------------------------------------------------------------ Learn Windows Azure Live! Tuesday, Dec 13, 2011 Microsoft is holding a special Learn Windows Azure training event for developers. It will provide a great way to learn Windows Azure and what it provides. You can attend the event by watching it streamed LIVE online. Learn more at http://p.sf.net/sfu/ms-windowsazure _______________________________________________ OpenBabel-Devel mailing list OpenBabel-Devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-devel
