Thank you Greg. I have to verify your example with my instance of RDKit, maybe my old version behaves differently. Which part of the code (C++ or Python part) is responsible for calculating coordinates?
On Thu, Aug 21, 2014 at 4:40 AM, Greg Landrum <[email protected]> wrote: > I can't seem to reproduce the problem. > > Here's an example showing that the original atom coordinates are preserved > when Chem.AddHs is called with the addCoords argument: > > #-------------------------------- > > In [9]: mb=""" > ...: Mrv0541 08211405312D > ...: > ...: 6 6 0 0 0 0 999 V2000 > ...: -3.5578 0.8250 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 > 0 > ...: -4.2723 0.4125 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 > 0 > ...: -4.2723 -0.4125 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 > 0 > ...: -3.5578 -0.8250 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 > 0 > ...: -1.8268 -0.7219 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 > 0 > ...: -2.8433 0.4125 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 > 0 > ...: 1 2 1 0 0 0 0 > ...: 1 6 2 0 0 0 0 > ...: 2 3 2 0 0 0 0 > ...: 3 4 1 0 0 0 0 > ...: 4 5 2 0 0 0 0 > ...: 5 6 1 0 0 0 0 > ...: M END > ...: """ > > In [10]: m = Chem.MolFromMolBlock(mb) > > In [11]: nm = Chem.AddHs(m,addCoords=True) > > In [12]: print Chem.MolToMolBlock(nm) > > RDKit 2D > > 12 12 0 0 0 0 0 0 0 0999 V2000 > -3.5578 0.8250 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -4.2723 0.4125 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -4.2723 -0.4125 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -3.5578 -0.8250 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -1.8268 -0.7219 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -2.8433 0.4125 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -3.5578 1.9250 0.0000 H 0 0 0 0 0 1 0 0 0 0 0 0 > -5.2249 0.9625 0.0000 H 0 0 0 0 0 1 0 0 0 0 0 0 > -5.2249 -0.9625 0.0000 H 0 0 0 0 0 1 0 0 0 0 0 0 > -3.8108 -1.8955 0.0000 H 0 0 0 0 0 1 0 0 0 0 0 0 > -0.8095 -1.1404 0.0000 H 0 0 0 0 0 1 0 0 0 0 0 0 > -2.1500 1.2665 0.0000 H 0 0 0 0 0 1 0 0 0 0 0 0 > 1 2 2 0 > 1 6 1 0 > 2 3 1 0 > 3 4 2 0 > 4 5 1 0 > 5 6 2 0 > 1 7 1 0 > 2 8 1 0 > 3 9 1 0 > 4 10 1 0 > 5 11 1 0 > 6 12 1 0 > M END > > #-------------------------------- > > Note that addCoords does add the Hs in 3D: > > #-------------------------------- > > In [5]: print Chem.MolToMolBlock(m) > > RDKit 2D > > 2 1 0 0 0 0 0 0 0 0999 V2000 > -0.7500 0.0000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > 0.7500 -0.0000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > 1 2 1 0 > M END > > In [7]: nm=Chem.AddHs(m,addCoords=True) > > In [8]: print Chem.MolToMolBlock(nm) > > RDKit 2D > > 8 7 0 0 0 0 0 0 0 0999 V2000 > -0.7500 0.0000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > 0.7500 -0.0000 0.0000 C 0 0 0 0 0 0 0 0 0 0 0 0 > -1.1167 0.0000 1.0371 H 0 0 0 0 0 1 0 0 0 0 0 0 > -1.1167 0.8981 -0.5185 H 0 0 0 0 0 1 0 0 0 0 0 0 > -1.1167 -0.8981 -0.5185 H 0 0 0 0 0 1 0 0 0 0 0 0 > 1.1167 -0.0000 1.0371 H 0 0 0 0 0 1 0 0 0 0 0 0 > 1.1167 -0.8981 -0.5185 H 0 0 0 0 0 1 0 0 0 0 0 0 > 1.1167 0.8981 -0.5185 H 0 0 0 0 0 1 0 0 0 0 0 0 > 1 2 1 0 > 1 3 1 0 > 1 4 1 0 > 1 5 1 0 > 2 6 1 0 > 2 7 1 0 > 2 8 1 0 > M END > > #-------------------------------- > > It is probably feasible to modify the code to add Hs in 2D when the > conformer is 2D, but it's likely a non-trivial amount of work and, as > Dimitri pointed out, there are many situations where the results aren't > going to look particularly good. > > -greg > > > > On Wed, Aug 20, 2014 at 7:34 PM, Greg Landrum <[email protected]> > wrote: >> >> >> >> On Wednesday, August 20, 2014, Michał Nowotka <[email protected]> wrote: >>> >>> >>> I've just checked and actually adding this flag doesn't solve my problem. >>> It seems that adding this flag actually calls Compute2dCoords (at >>> least this is how it looks like), because all coordinates are computed >>> from scratch, for example if my compound was translated by some vector >>> on 2D plane, after executing addHs it will be centered, if the >>> compound had one atom dragged unusually, it will be aligned back. >>> >>> The only difference between this flag and calling Compute2dCoords is >>> that H bonds are longer, which actually looks worse... >>> >> >> Hmmm, that doesn't sound like what is supposed to happen. I think that it >> should just add coords for the Hs that it adds. I will take a look tomorrow >> morning to see what is going on. There may be something odd for >> 2d conformers, but I really don't think it should redo the whole >> coordinate generation. >> >> -greg > > ------------------------------------------------------------------------------ Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Rdkit-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
