Dear modellers,
we use Modeller6 and we have some problems restraining beta-sheets in our model. The student working on this project sent an -email to the forum a few days ago and since then we have tried many different things to try to analyze the problem so we send more precise questions.
We tried to restrain beta strands in our model, using the RESTRAINT_TYPE = 'SHEET' or RESTRAINT_TYPE = 'STRAND' . All our tests were unsuccessful. Last thing we tried was to force beta strands in the model to be longer than they are without special restraints. That means, in the previous model, there are beta strands (they exist in the template structure) without adding other special restraints in the top script. I thus think it should be possible to make them slightly longer. But if I try to apply special restraints to increase their length, the strands completely disappear in the models.
The .top file is OK, modeller can read it, there are no error messages in the .log file, the .rsr file contains additionnal restraints (dihedral pairs) with regards to the same calculation without any restraint. The .V files are also different with or without the special restraints. So everything seems to be OK, but the models never contain these additionnal beta strands. The longer the strands we try to restrain, the more violations (in the log file) there are but their magnitude is not significantly higher than when I apply an 'alpha' restraint (which works perfectly well, i.e. it appears in the model whatever the region we try to introduce it in). To force strand (or even beta-sheets) never worked. So it looks like restraining strands and beta-sheets is much more difficult than alpha-helices, which we can introduce almost anywhere in the model.
Here are our questions : Is it normal to be able to force an alpa helix wherever you want and not a beta, and if yes why ? Are the violations due to beta strands too penalizing ? Shall we modify the optimization schedule and if yes, which parameters should we change ? Is it possible to exclude block of atoms from the optimization procedure of the objective function ?
Thanks,
Nathalie Reuter
-- Nathalie REUTER, PhD U.410 INSERM Faculte de medecine Xavier Bichat 16, rue Henri Huchard BP 416 75870 PARIS Cedex 18 tel : (33) 01.44.85.61.32 fax : (33) 01.42.28.87.65 -- Nathalie REUTER, PhD U.410 INSERM Faculte de medecine Xavier Bichat 16, rue Henri Huchard BP 416 75870 PARIS Cedex 18 tel : (33) 01.44.85.61.32 fax : (33) 01.42.28.87.65
If you do not delete the homology derived restraints, they will clash with the new ones. This might be a problem, but additionally there always can be some other problems. If you send your example top, and coordinate file we will take a look.
Andras
Reuter Nathalie wrote: > > Dear modellers, > > we use Modeller6 and we have some problems restraining beta-sheets in > our model. The student working on this project sent an -email to the > forum a few days ago and since then we have tried many different things > to try to analyze the problem so we send more precise questions. > > We tried to restrain beta strands in our model, using the > RESTRAINT_TYPE = 'SHEET' or RESTRAINT_TYPE = 'STRAND' . All our tests > were unsuccessful. Last thing we tried was to force beta strands in the > model to be longer than they are without special restraints. That > means, in the previous model, there are beta strands (they exist in the > template structure) without adding other special restraints in the top > script. I thus think it should be possible to make them slightly longer. > But if I try to apply special restraints to increase their length, the > strands completely disappear in the models. > > The .top file is OK, modeller can read it, there are no error messages > in the .log file, the .rsr file contains additionnal restraints > (dihedral pairs) with regards to the same calculation without any > restraint. The .V files are also different with or without the special > restraints. So everything seems to be OK, but the models never contain > these additionnal beta strands. The longer the strands we try to > restrain, the more violations (in the log file) there are but their > magnitude is not significantly higher than when I apply an 'alpha' > restraint (which works perfectly well, i.e. it appears in the model > whatever the region we try to introduce it in). To force strand (or even > beta-sheets) never worked. So it looks like restraining strands and > beta-sheets is much more difficult than alpha-helices, which we can > introduce almost anywhere in the model. > > Here are our questions : > Is it normal to be able to force an alpa helix wherever you want and not > a beta, and if yes why ? > Are the violations due to beta strands too penalizing ? > Shall we modify the optimization schedule and if yes, which parameters > should we change ? > Is it possible to exclude block of atoms from the optimization procedure > of the objective function ? > > Thanks, > > Nathalie Reuter > > -- > Nathalie REUTER, PhD > U.410 INSERM > Faculte de medecine Xavier Bichat > 16, rue Henri Huchard > BP 416 > 75870 PARIS Cedex 18 > tel : (33) 01.44.85.61.32 > fax : (33) 01.42.28.87.65 > -- > Nathalie REUTER, PhD > U.410 INSERM > Faculte de medecine Xavier Bichat > 16, rue Henri Huchard > BP 416 > 75870 PARIS Cedex 18 > tel : (33) 01.44.85.61.32 > fax : (33) 01.42.28.87.65
participants (2)
-
Andras Fiser -
Reuter Nathalie