I would like to know, apart from Maxsprout and sabbac, which other methods do you use to convert a protein model that only contains alpha carbons to an all-atom representation, and if you can supply with the input files (or at least just the first lines) of the input structures you use.
There are two steps in the process you've described: 1.) reconstruction of protein backbone from C-alpha trace 2.) reconstruction of full atom model from protein backbone. I would use any of the two mentioned (maxsprout or sabbac, I don't know of other methods that actually have been implemented as a working software) to generate protein backbone (or if the full structure is generated, I would discard side chain coordinates), and then use SCWRL to generate full structure. I think structures optimized with SCWRL are much better (if you actually care about such level of detail).
Regarding building a full backbone model from CA trace, I absolutely wouldn't use Maxsprout, it's over two decades old and was designed using a handful of structures. Here are some currently-available programs ranked by accuracy: BBQ > SABBAC > PULCHRA > Maxsprout (I don't have a reference for this, it's my own benchmarking).
Some of these will also position side chains but I have no experience there and it's probably better to use a specific tool such as that recommended by Pawel.
The input files are all standard PDB, i.e. fixed width. Download any structure from RCSB and make sure yours lines up (there's a fair amount of documentation about the format too).
Update: A paper describing a novel method I helped develop for backbone reconstruction and the benchmarking I allude to has since been published here.
Probably the best you can get is a program like CNS/Xplor that can rebuild the side chains and perform a slight energy minimization on the entire structure to get things going. SCWRL might do the trick though if you just want to add side chains and then use the structure for some other purpose like MD.
Maybe you can try to rebuild a model from your Ca template with Modeller.
Actually you can use the catoallatom application from Rosetta.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
I have never used modeller but people told me it is very complicated to use. Could you point to some tutorial?
Their own webpage: http://www.salilab.org/modeller/tutorial/