Re: [IMP-dev] Q) How to correlate atoms and residue in IMP molecular hierarchy?

[Daniel Russel <>]

MolecularHierarchy inherits from Hierachy, so you can do operations  
like get_parent and get_child. And by default, atoms are children of  
residues. So you can do atomdecorator.get_parent().get_index() to find  
which residue a given atom belongs to. Does that answer things?

On Jan 7, 2009, at 7:34 PM, Seung Joong Kim wrote:

> Hi all,
>
> This can be a very simple question, but I'm just wondering how to  
> correlate atoms and a residue in IMP molecular hierarchy.
>
> First of all, I get "IMP::core::MolecularHierarchyDecorator mp" by  
> calling read_pdb function defined in IMP modeller module.  It simply  
> reads contents of a pdb file and generates the molecular hierarchy.
>
>   mp = IMP.modeller.read_pdb('single_protein.pdb', m)
>
>
> Then, I load particles by calling "molecular_hierarchy_get_by_type"  
> function as follows.
>
>   ps_ = IMP::core::molecular_hierarchy_get_by_type(mp_,  
> IMP::core::MolecularHierarchyDecorator::ATOM);
>   IMP::core::AtomDecorator ad =  
> IMP::core::AtomDecorator::cast(ps_[i]);
>   IMP::core::AtomType at = ad.get_type()
>
>   residues = IMP.core.molecular_hierarchy_get_by_type(mp,  
> IMP.core.MolecularHierarchyDecorator.RESIDUE);
>   IMP::core::ResidueDecorator rd =  
> IMP::core::ResidueDecorator::cast(residues_[i]);
>   IMP::core::ResidueType rt = rd.get_type();
>
> By doing this, I get only independent particle arrays of atom and  
> residue, without any relation between them.
>
>
> What I want to know is a piece of linked information between atom  
> index and residue index, such that a Nitrogen atom (atom index 458)  
> is positioned at a LYS residue (residue index 56).
>
> ATOM    458  N   LYS    56      38.397  16.302  23.700  1.00  
> 18.45         N
> Is there any simple function defined in MolecularHierarchyDecorator  
> class for this purpose?  (Input : atom index & Output : residue  
> index, and vice versa.)  Or is there any other simple way to do this?
>
> Thanks a lot!
> Seung Joong Kim
> _______________________________________________
> IMP-dev mailing list
>
> https://salilab.org/mailman/listinfo/imp-dev