What about just eliminating the divide by zero? Anyone have any idea where this is located in eng1_mm_tripos52.cpp? I guess I'll look through the code at some point this weekend. I'm sure at some point i can figure it out. Thanks for the heads up Geoff. -Donald (Fri, Dec 15, 2006 at 06:38:01PM -0500) Geoffrey Hutchison <geoff at geoffhutchison.net>: > > On Dec 15, 2006, at 4:41 PM, Donald Ephraim Curtis wrote: > > >Before I add the duplicate atom coordinate "check" i wanted to make > >sure > >we couldn't fix this problem in the libghemical tools. > > I think this should ideally be fixed in the underlying libghemical. > Basically, there's a divide by zero and that kills everything. You > can see the same effect if you import a SMILES string (which will > have 0.0 for all coordinates). > > So I think libghemical needs a quick check for coordinates which are > too close. (For example, sort by z-axis, and check for distances > under 0.4Ang / 40pm.) > > Cheers, > -Geoff