Hello,

I am trying to cluster protein 3D structural motifs in some form of feature space. To do this, I need a 'similarity metric' of sorts for each motif I look at (for our purposes, a 'motif' is defined as simply a subset of the whole protein structure; e.g. PDB file containing only the residues within 10Å of a given residue-of-interest; say, 6 residues from the original protein).

I know that I can get RMSD from 'aligning' two protein structures with a program such as pymol; but I have an extra 'known variable' that I would like to use. When comparing motif A with motif B, I will always have a known 'point of comparison', e.g. residue S123 on protein A and residue S456 on protein B. I need these two points to be overlayed, and set to be immovable relative to eachother; and THEN find the optimal 'fit' or structural alignment between the two structures (I would imagine this would be simply rotating the structures around relative to eachother; while maintaining the common anchor point of the two set residues S123 and S456 such that they are always on top of eachother).

I have looked through the docs for various pymol functions for superimposing structures but cannot find a way to 'enforce' two positions to be stuck together.

Thanks for any advice in advance!