I forgot my important question in all my "this seems to be working but the test says it doesn't"-drama:

How do I specify disulphide bonds for MR_rosetta? I know it keeps them if they are in the input model but disulphides are generally poorly conserved, especially in the %identity range where mr_rosetta is the most useful. 

According to the changelog this feature should be available in the latest dev.

Cheers,

Morten
--
Morten K Grøftehauge, PhD