## jmol-users

 Re: [Jmol-users] If Jmol ever has to artifically scale substructure views, how should/would it do so? From: - 2008-07-25 14:49:55 Attachments: Message as HTML ```Again, Bob, many thanks for taking the time. For method2 (the q-map), I will have to do the requisite background reading - thanks for the directional signal. However, method1 is closer (I think) to what I need, as the following should make clear (I hope!). What I would do is derive the initial set of "driver" vectors by an algorithm over the pair of linear DSSP or STRIDE segments Sa and Sb corresponding to the tertiary substructures A and B. (One can easily image how such an algorithm would work: 1) first arrange the helices and strands of the smaller DSSP/STRIDE segment so that they are (linearly) "on-center" with the corresponding helices and strands of the larger DSSP/STRIDE segment; 2) then work on each helix of the smaller DSSP/STRIDE segment so that its start and end residues coincide (linearly) with the start and end residues of the corresponding helix in the larger DSSP/STRIDE segment; then, for each helix in the smaller segment, space its residues evenly between the new positions of its start and end residues; 3) do (2) for strands as well as helices. 4) then "space" residues in the intervening "turns" or :"loops" joining the strands and helices. Once (1-4) are completed, it would be trivial (I think) to derive the "driver" vectors required for the JMol deformation. Reason why I prefer this approach, at least initially, is that I can: do a visual superimposition to see which residues of the larger (unscaled) substructure line up closely with which residues of the smaller (scaled) structure, followed up by a "calculated" superimposition to make a table of corresponding residues. And then, this table of corresponding residues might provide a way to refine and elaborate the original "sparse" primary structure patterns that led to the initial identification of substructrues A and B in the first place, i.e. make these primary structure patterns less "sparse" and therefore increase their usefulness as probes back into the PDB. Anyway, thanks very much again for providing direction here. djh ```