Bo Yang wrote: > I try to do a search for potentially related sequences of known > structure by using the "build_profile.py" from the "command" section of > Modeller9v1. But the program complaines that "Cannot open file > pdb95.fsa: no such files". I search the Modeller9v1 package, and can't > find the file named as "pdb95.fsa".

The pdb95.fsa file is in the examples/commands directory - the same directory as the build_profile.py script.

> Do you know where can I find the "pdb95.fsa" file? Or is there an > alternative file that can be used as the "pdb95.fsa" in the program > package? I notice that there is a file called "pdball.pir" from the > download section of Modeller. Is this one can be used as a database? If > so, how can I change the file format to .fsa?

pdb95.fsa is just all of the PDB sequences, filtered at 95% sequence identity to remove redundancy. The pdball.pir download from our website is the same thing. You can convert from PIR to FASTA format using alignment.read() and alignment.write(), but there is no need, because profile.read() can also read PIR format databases - just edit build_profile.py and change seq_database_format from 'FASTA' to 'PIR'.

Ben Webb, Modeller Caretaker