[Biopython-dev] test_SeqIO_online failure

[Phillip Garland]

The fasta formatted record is fine, the problem seems to come after
requesting and reading the genbank-formatted record for the protein
with GI:16130152.

It looks like the record was modified a few days ago:

LOCUS       NP_416719                367 aa            linear   CON 24-MAY-2013

and ends with

CONTIG      join(WP_000865568.1:1..367)\n//\n\n'

instead of

ORIGIN and the sequence data.

Is this a problem with the genbank record that should be reported to
NCBI, or is SeqIO supposed to handle the record as it is by fetching
the sequence from the linked contig, or is the test doing the wrong
thing by using rettype="gb" instead of rettype="gbwithparts"?

Here's the test output:

pgarland at cradle:~/Hacking/Source/Biology/biopython/Tests$ python
run_tests.py test_SeqIO_online.py
Python version: 2.7.5 (default, May 20 2013, 11:51:12)
[GCC 4.7.3]
Operating system: posix linux2
test_SeqIO_online ... FAIL
======================================================================
FAIL: test_protein_16130152 (test_SeqIO_online.EntrezTests)
Bio.Entrez.efetch(protein, 16130152, ...)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/pgarland/Hacking/Source/Biology/biopython/Tests/test_SeqIO_online.py",
line 77, in <lambda>
    method = lambda x : x.simple(d, f, e, l, c)
  File "/home/pgarland/Hacking/Source/Biology/biopython/Tests/test_SeqIO_online.py",
line 65, in simple
    self.assertEqual(seguid(record.seq), checksum)
AssertionError: 'NT/aFiTXyD/7KixizZ9sq2FcniU' != 'fCjcjMFeGIrilHAn6h+yju267lg'

----------------------------------------------------------------------
Ran 1 test in 10.010 seconds

FAILED (failures = 1)

~Phillip