> > > Miguel wrote: > >>>>I have reproduced the problems with 1PPA. I am puzzled. I will >>>>investigate. >>>> >>> >>>a hint may be, there is no secondary structure defined >>>in 1ppa if I inserted >>>TURN 1 T1 SER 1 SER 1 The dummy structure >>>into 1ppa.pdb >>>no 'hbonds' where shown. >> >> >> This was a good hint, but I did not grasp the significance when I first >> read the message. >> >> The problem is that the built-in algorithm for calculating secondary >> structures is dependant upon hbond interactions. > > Hmm. Are you sure? What's that about using angles > instead of hydrogen bonds?
I saw the message exchange with Tim. When I worked on it (about 2 years ago) ... porting the psuedo-DSSP stuff from RasMol it was still dependent upon hbonds. I did not see Tim's comment and did not realize that the hbonds were no longer being used at all in secondary structure identification. > Eclipse reports only two access points to > calcHydrogenBonds: via AminoPolymer.calculateStructures() > and in HSticks.setSize(). But I can't find any use in > calculateStructures() for the information returned by > calcHbonds. If that's true, and the calculation really > isn't used, I'm just wondering if it was supposed to be > commented out. That would be good ... simplify things quite a bit. > That would leave setSize(), which isn't > ideal either, but isn't that much of a concern. We could > just use if(size > 0) calcHbonds().... I think that we will have to do more than that ... because if someone defines their own hbonds then they try to set the size they would not the 'autocalc' hbonds to appear. Because of this I still think that we need to introduce 'hbond calculate' or equiv. The important result is that it isn't required for secondary structure calculation. > Ah, one problem is that there are several methods: > > calcHbonds() > only called by HSticks > calls Mmset.calcHydrogenBonds() > > Mmset.calcHydrogenBonds() > called in Frame.freeze() when structures not present > called by calcHbonds() > calls ModelManager.calcHydrogenBonds() > > ModelManager.calcHydrogenBonds() > called by Mmset.calcHydrogenBonds() > > AminoPolymer.calcHydrogenBonds() > called by ModelManager.calcHydrogenBonds() > > NucleicPolymer.calcHydrogenBonds() > called by ModelManager.calcHydrogenBonds() > > > A bit of a nettle, I think. Still, I don't think any of it is needed for > structure, is it? I think Tim said no. You didn't see anything. I will also take a look. >> So, if there were any structures in the file, then the internal hbond >> calculations were not performed. >> >> However, if there were no secondary structures defined in the input file >> then the internal hbond calculations *do* get applied, causing undesired >> interactions. > > The problem only arises with files with no secondary structure then, > right? We > know that calcHbonds() is being run unnecessarily when hbonds are turned > OFF due > to restrict(). We also know that their diameters are immediately set to 0. > I > don't think that's the problem. It must be that calculateStructure() is > also > being run. But calculateStructure() doesn't seem to need hydrogen bonds. > > I suggest just commenting out that calcHbonds() in calculateStructure() > and > seeing if the problem goes away. It could be just that simple, and we > could > worry about fundamentals later. Yes. I will try this. As I said above, I still think there is an issue with files that do not have structures defined. If you say 'connect hbond ...; hbond 0.1' I think that the calculated hbonds are going to show up. I'll report back. Miguel ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid0944&bid$1720&dat1642 _______________________________________________ Jmol-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-developers
