Tamas Hegedus wrote: > I have an alignment file with 7 proteins (including the 'sequence'). > I got the following error: > > _modeller.error: crsplin_223E> Internal error: array too small: MAXSPL > current maximum, current need: 400 488 > > My guess: you have some hard coded buffer size, and I have too many and > large (?) proteins in the alignment file...

Yes, there is a hard-coded limit on the number of spline points in Modeller 8, and your templates are sufficiently different that Modeller cannot represent the multi-Gaussian PDFs as splines. You can work around this by setting spline_on_site to False, although if your templates are really very different you would be better advised to improve your alignment.

> If I have 1XMI in the alignment file, I have the following error: > > _modeller.error: rdpir___648E> Alignment sequence not found in PDB file: > 1 ./pdb/1XMI.pdb > > I checked for extra spaces, numbering, sequence, but I do not find the > problem...

You have specified the chain IDs as '.' in your alignment, which instructs Modeller to check your sequence against the PDB to determine the real chain IDs. This will only work if your sequence matches exactly, which it does not. Replace each '.' with the real chain ID (A) and run your script again, and then Modeller will show you where the mismatch occurs.

Ben Webb, Modeller Caretaker