[BioPython] PSIBlastParser error

[Jer-Yee John Chuang]

Hi,

I am some difficulty with the NCBIStandalone.PSIBlastParser.  After
running blastpgp, I use the parser and get the error message:

chuang at Poggi ~/blast/workbench
$ python psiBlastHomologues.py ./275proteins_allEnteroBact/allEntero
./275prote
ins_allEnteroBact/all275seeds 1 100 1e-3 0 1
Traceback (most recent call last):
  File "psiBlastHomologues.py", line 89, in ?
    myRecord= myIterator.next()
  File "NCBIStandalone.py", line 1342, in next
  File "NCBIStandalone.py", line 581, in parse
  File "NCBIStandalone.py", line 96, in feed
  File "NCBIStandalone.py", line 154, in _scan_rounds
  File "NCBIStandalone.py", line 279, in _scan_descriptions
  File "c:\python23\lib\site-packages\Bio\ParserSupport.py", line 340, in
read_a
nd_call_until
    method(line)
  File "NCBIStandalone.py", line 647, in description
  File "NCBIStandalone.py", line 704, in _parse
  File "NCBIStandalone.py", line 1612, in _safe_int
ValueError: invalid literal for float(): round


I've traced this problem to the following line in the blast output:
"Sequences not found previously or not previously below threshold:"   It
appears that the parser complains when this line is not followed by any
sequences AND there is another round afterwards.  This is kind of odd,
since I thought that when there are no sequences following that line,
psiblast should have converged.  However, this is not the case, as I have
additional rounds (i.e. iterations) afterwards.  I can forward the actual
psiblast output file if it helps.

Thank you for your time.

Sincerely,
John Chuang