On 11/25/2010 07:45 AM, flavio seixas wrote: > Hi, I am modellig a structure of an hemoglobin with the four heme > groups bonded to oxygen. > > To do the alignment I refer to heme as "h" and oxygen as "o". > > The template model is the oxy hemoglobin at 1.25 Angstrons resolution > Pdb id: 2DN1. > > The final model has heme and oxygen, but the iron configuration is > OUT of the plane of porphyrin ring toward proxymal His which is > bonded to iron. > > This configuration is different as observed in oxygen ligated form of > hemoglobin in which the iron is in the SAME plane of the porphyrin > ring.

If you use one-letter codes for ligands (or any other non-protein residue) you will need to define the connectivity between them, just as you would need to do in a molecular mechanics package. Modeller will not magically connect them and realize that your iron should be in the same plane as the oxygens.

If, however, you use the BLK residue type ('.' one letter code) for non-protein residues, Modeller will keep each residue rigid (by restraining internal distances) and also do its best to preserve the same residue-residue distances it finds in the template in your output model. So this is probably what you want in your case.

Ben Webb, Modeller Caretaker