[Bioperl-l] Help with Bio::SeqIO

Jay Hannah jay at jays.net
Mon Nov 5 15:14:17 UTC 2007


On Nov 4, 2007, at 7:08 PM, Chris Fields wrote:
> To the other devs: shouldn't -nosort be the default behavior when the
> split location is a 'join'?

I certainly think so.

> In other words, should spliced_seq() be
> modified to take into account the split location type when returning
> sequence?  GB/EMBL/DDBJ rel. notes indicate a 'join' explicitly
> indicates the order of the sequences is important when joined
> together; the current behavior is more like that for 'order'.

I don't see any value to the sorting algorithm. All tests invoke - 
nosort => 1 (except a phase test where nosort doesn't matter anyway).  
In my limited experience the sorting only serves to break real-world  
splicing.

If there is no valid use then we can remove ~20 lines from  
SeqFeatureI.pm circa line 505. If there is a valid use and someone  
would be so kind as to educate me I'd be happy to add tests which  
demonstrate them.  :)

P.S.  CSHL is neato. I plan on understanding some of this stuff some  
day.  :)

j
http://www.bioperl.org/wiki/User:Jhannah






More information about the Bioperl-l mailing list