On 24/04/2008, Greg Landrum <[email protected]> wrote: > Hi Noel, > > > On Thu, Apr 24, 2008 at 12:03 PM, Noel O'Boyle <[email protected]> wrote: > > I'm a bit confused by how RDKit calculates the 2D coordinates. I use > > AllChem.Compute2DCoords(mol, clearConfs = False) to calculate the 2D > > coordinates. > > > > Given an SD file containing a 2D depiction, RDKit appears to use those > > coordinates rather than generating its own. Can you confirm that this > > is the case, as this is a bit unexpected? > > > Take a look at the documentation for AllChem.Compute2DCoords > (accessible in Python via: "help AllChem.Compute2DCoords). > By setting clearConfs=False you have told the code to keep the > existing conformation (the one read from the SD file) and to add the > 2D conformation as a new one. If you had left that argument in its > default value (True), then the existing conformation(s) would have > been replaced with the new 2D one. > > an example: > [5]>>> m = Chem.MolFromMolFile('foo.mol') > [6]>>> m.GetNumConformers() > Out[6] 1 > [7]>>> AllChem.Compute2DCoords(m,clearConfs=False) > Out[7] 1 > [8]>>> m.GetNumConformers() > Out[8] 2 > [9]>>> AllChem.Compute2DCoords(m) > Out[9] 0 > [10]>>> m.GetNumConformers() > Out[10] 1
Right. In my initial testing I only used SMILES strings, which of course don't have any conformers to begin with.
