Hi Nicolas

It does happen sometimes although rarely, that "knots" show up. Like in your case, when a very artificial setup is forced.

1., it would be better to have overlaps among the connected segments, that will provide info about the relative orientation of each segment

2., (hit-and-miss approach) generate several models ("set FINAL_MODEL= 5" (20)?) . And eliminate those ones that goes wrong. You can also introduce an even more loosely defined starting conformation by increasing the deviation (e.g "SET DEVIATION= 5.0", check the manual, RANDOMIZE_XYZ command )

Andras

> Dear users, > > I made a quite unusual model, with three structures as templates, linked > by tethers which (presumably) do not present any secondary structures: > > >P1;tempA > structureX:tempA:1::10::::: > aaaaaaaaaa------------------------------ > > >P1;tempB > structureX:tempB:1::10::::: > ---------------bbbbbbbbbb--------------- > > >P1;tempC > structureX:tempC:1::10::::: > ------------------------------cccccccccc > > >P1;model > sequence:model:::::::: > aaaaaaaaaaTTTTTbbbbbbbbbbTTTTTcccccccccc > > Everything's fine, except that in the final model one tether cross > directly through the domain C! It is already the case in the file .ini > and that seems related to the initial relative positions of the three > templates. Is it supposed to occur? Any idea how to fix that? > > Regards, > > -- > Dr Nicolas Le Novère http://www-alt.pasteur.fr/~lenov > Dpt of Zoology, Univ of Cambridge, Downing street, Cambridge CB2 3EJ, UK > tel: +44 1223 336623 fax: +44 1223 336676 e-mail: nl223@cus.cam.ac.uk