Eric wrote: > I could well be missing something(s), but I think a simple > straightforward > solution to the handling of altLoc's should be possible.
I am eager to hear suggestions :-) > The PRIMARY objectives as I see it are: > > 1. Enable users to select altLocs with complete flexibility. A user should > be able to select and display concurrently e.g. altLoc A at residue 50, > altLoc B at residue 100, altLoc C at residue 150, and altLoc A at residue > 200; or any arbitrary permutation. > > 2. It should also be easy to select all altLoc A's as a group; or all > altLoc B's as a group, etc. > > 3. Inappropriate covalent bonds must be avoided (no spaghetti, basket > weave, or birds nests). These are all desirable goals. #3 says to avoid 'spagetti' bonding. You did not say anything about whether all the bonds must be in place. That is, if we swap run A for run B, do we expect the bonding from the common ' ' atoms to the altLoc runs to stay in place. Clearly that is desirable, but it quickly pushes us back to calculating all possible combinations ... particularly if there can be connectivity between different runs. > [Chime can do none of 1, 2, or 3. It shows all altLocs as spaghetti.] Jmol does not read the altLoc atoms. > These objectives seem to me to contribute to simplification of > IMPLEMENTATION as follows: > > 4. It is fine to handle, internally within Jmol, all altLoc A's as one > "model" (but using different selection/display syntax than for NMR > MODEL/ENDMDL models), all altLoc B's as a different "model", etc. > Permutations can be handled by selection/display commands. There is no > need > to have all possible permutations listed internally somehow. OK > (I believe > that a long time ago I may have objected to this concept. If so, I have > changed my mind.) Good. > 5. Handling A's as a distinct model from B's, distinct from C's, and > prohibiting bonds between altLoc-models takes care of spaghetti. Correct. But, I strongly suspect that it will also lead to 'missing connectivity' unless one is looking at all of the 'A's or all of the 'B's. > RECOMMENDATIONS: > > 6. It will often be sufficient to display all altLoc A's, then hide A's > and > display B's, then return to A's, so as to toggle back and forth. Item #2 > above enables this mode of "toggling" display. Of course, if the user > wants > to toggle arbitrary permutations, they can do so using the > residue-specific > altLoc selection as in #1 above. Correct. It seems to me that the syntax for doing this might be complicated. As I understand it, you are generally going to be working with runs of sequential altLocs, not individual atoms. We may need to add some query commands that expand a given altLoc atom to its neighbors in the same run. > 7. I am in favor having all altLocs displayed concurrently in the initial > default view of Jmol. Jmol does not show multiple models by default, so I would tend to say that it should now show multiple altLocs by default ... but we'll see how it turns out. I suspect that this is a relatively minor point relative to the rest of the work. > This avoids having people be unaware of the altLocs. > On Jmol's menu could be a new AltLoc's section. The default would be "All > altLocs", with other options being A, or B, or C (listing only those that > actually exist in the loaded model). (Display of permutations would need > to > be handled with custom scripting.) Another menu item could be "animate > altLocs" which would show each "model" (A, B, C) in turn for half a second > or so. These menu options would handle 95% of people's needs, I suspect. Let's keep talking. Miguel ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642 _______________________________________________ Jmol-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-users
