On 06/09/2015 08:32 AM, Kolmus, Elizabeth K. wrote: > I am attempting to model a target sequence from multiple templates. When > I check the alignment, I'm getting an error I can't resolve either on my > own or by consulting previous entries on the mailing list. It appears > that the program is concatenating the end of one chain directly onto the > beginning of the other and ignoring a gap.
Modeller doesn't do anything special. It reads the PDB file strictly sequentially. Every residue in the PDB file needs to also be in the alignment file, in the same order. Gaps (and chain breaks) are used to align the template sequence with the target - they have no effect on the matching between alignment file and PDB.
> "This command can raise ... a > SequenceMismatchErrorif a 'PIR' sequence does not match that read from > PDB (when an empty range is given)" - does this mean that automodel will > not run on the alignment? What is the appropriate fix?
The fix is simple - your alignment sequence must match the primary sequence from the PDB. Looking at 3TNP from PDB, it does not match your alignment.
> I suspect I may also have trouble down the line with matching the chains > together, given that 3TNP is formatted as 1-2-1-2 and 3J4Q is formatted > as 3-1-1-2-2. Will this be an issue, and if so, what is the workaround?
Modeller doesn't care what order the chains are in in your PDB files, as long as you put them in the same order in the alignment file. All the information it uses to build the model is structural. If you want to change the order, edit the input PDB files accordingly.
Ben Webb, Modeller Caretaker