[Bioperl-l] t/entrezgene, t/protgraph and t/FeatureIO

Chris Fields cjfields at uiuc.edu
Mon Aug 7 14:35:25 UTC 2006


On Aug 7, 2006, at 9:04 AM, Sendu Bala wrote:

> Chris Fields wrote:
>> Brian already knows about the 'pseudohash' issue with protgraph.
>
> With entrezgene you mean?

Yes, sorry.  Lack of coffee.

> Do you get the same test fails for protgraph on Windows that I get  
> using
> linux?

The tests all pass on Mac OS X; Ill try Windows today, but if memory  
serves there were problems with this a few months back

Okay, found it on the gmane lists:

http://thread.gmane.org/gmane.comp.lang.perl.bio.general/11137/ 
focus=11311

This part of the thread details the issues.  Linux and WinXP fails  
similar tests, Mac OS X (Tiger) passes (Heikki and I ran the tests).   
Never resolved.

My feeling: could be OS-specific, could be Clone.pm, could be perl  
version.

>
>> I think it's a weird ref call or construct that tripping the perl
>> warnings.  I don't think perl 5.6 does this; the pseudohash warning
>> was added in perl 5.8.
>>
>> I really wouldn't worry about the other tests throwing warnings.  You
>> have your hands full with Taxonomy now; focus on getting that up and
>> running before you move on to other ventures.
>
> As far as I'm concerned I'm more or less finished with Taxonomy. I'm
> just waiting for your confirmed 'vote' on what species() should  
> return,
> and at least one other vote.
>
> Might also be waiting to find out how get_lca should really work as
> well. But in any case, I'm just waiting, which isn't that hard ;)
>
> FYI I'm actually moving on to the previously mooted hmmpfam plugin for
> SearchIO.

Cool!  The only problem I saw with the way you wanted to go about it  
was you would have to preparse the report since everything is  
reported as 1 hsp/hit.  They usually aren't long, so that shouldn't  
be a problem.

Christopher Fields
Postdoctoral Researcher
Lab of Dr. Robert Switzer
Dept of Biochemistry
University of Illinois Urbana-Champaign






More information about the Bioperl-l mailing list