[MMTK] comparing modes from different universes

[Edvin Fuglebakk]

Hi

I am comparing normal modes between aligned structures.

I have superimposed structures with coordinates for the superimposition
coded in different pdb-files (one file for each structure).

Can I rely on MMTK to preserve these structures relative orientation to
each other when loaded into different universes so that the directions
of the atomic displacements are comparable?

If not, is there a way to mask out some of the structures while
calculating the normal modes so that the modes calculated for one
structure will not be affected by the presence of the other structures
in the same universe.

The following test seems to work, but I would like assurance that it is
a reliable way to do it:

    #load the structures into different universes
    uni=InfiniteUniverse(CalphaForceField())
    uni.s1=Protein(pdbfile1, model='calpha')
    uni2=InfiniteUniverse(CalphaForceField())
    uni2.s2=Protein(pdbfile2, model='calpha')

    #load the structures into the same universe
    combinedUniverse=InfiniteUniverse(CalphaForceField())
    combinedUniverse.s1=Protein(pdbfile1, model='calpha')
    combinedUniverse.s2=Protein(pdbfile2, model='calpha')

    #test that c-alpha positions are the same in the different universes
for first chain of each protein
    for i in range(min(len(uni.s1[0]), len(uni2.s2[0]))):
        if uni.s1[0][i].position() !=
combinedUniverse.s1[0][i].position() or uni2.s2[0][i].position() !=
combinedUniverse.s2[0][i].position():
            print "position in combined universe does not match position
in other universes"


Thanks
-Edvin Fuglebakk