Dear all,

I'm modelling a protein with some additional NMR-restraints for hydrogens.

It does work fine for about half of the models, for the other half I get very high scores in the MODELLER OBJECTIVE FUNCTION. This is due to violations of the bond length potential caused by wrong connections between Ns and Hs of an ARG residue (see logfile below), for the models with the 'good' scores the connections are OK.

Any ideas what might mess up the N-H bonds?

thanks in advance,

ralf

logfile:

<snip> ------------------------------------------------------------------------ Feature 1 : Bond length potential List of the RVIOL violations larger than : 4.5000 #ICSR RESNO1/2 ATM1/2 INDATM1/2 FEAT restr viol rviol RESTR VIOL RVIOL 1 774 53R 53R NH1 HH11 772 773 3.27 1.00 2.27 89.01 1.00 2.27 89.01 2 775 53R 53R NH1 HH12 772 774 2.57 1.00 1.57 61.57 1.00 1.57 61.57 3 776 53R 53R NH2 HH21 775 776 2.57 1.00 1.57 61.51 1.00 1.57 61.57 4 777 53R 53R NH2 HH22 775 777 3.27 1.00 2.27 89.02 1.00 2.27 89.02 ------------------------------------------------------------------------ <snip>

======================================= Dr. Ralf Schmid Institute of Cell and Molecular Biology The University of Edinburgh King's Buildings Michael Swann Building 3.14 Edinburgh EH9 3JR