Structure prediction

Dear all,

I have been battling with a specific homology modelling for some time (I never had problems before though) and hope that you can help me further. Whichever version of homology modelling I use (3.5 and never hybrid one from the latest release), the sequence of the final model does not correspond to the sequence of the alignment I provid. It appears that a gap very early in the alignment is skipped and all following residues are hence shifted by 2 (only one more gap in ~450 res). For simplicity just the first few residues:

Dear users,
I'd tried to run 50 000 models from ab initio Rosetta protocol using the options bellow [1]. After running some models my ab initio crashed due to problems with eletric supplies. I reinitiated the ab initio protocol intendind to generate my 50 000 models and I guess: if I'm very interested to analyse the "inital models" (that ones that was generated until the ab initio crashed), how can I "recover" these information from my out file until its crashed?

HI,

I was wondering what criteria Rosetta uses to determine when it switches from centroid to full atom mode? Is it a set number of decoys, a target energy value, radius of gyration or something else?

Thanks

Mark

Is it possible to ask Rosetta Ab initio application to read in a fragment file containing fragments of variable length generated by myself? If so, would it also have any impact to later stage analysis?