[Bioperl-l] Distance between non-overlapping sequences in DNAStatistics

Chris Fields cjfields at illinois.edu
Mon Jan 25 13:27:26 UTC 2010


That works for me, just want to ensure we're DTRT.  I'll change it over.

chris

On Jan 24, 2010, at 11:58 PM, Jason Stajich wrote:

> It could also return -1 which is used as place holder for NA in other programs that generate distance matrices.
> -jason
> On Jan 24, 2010, at 9:35 PM, Chris Fields wrote:
> 
>> Just a quick question for those using DNAStatistics.  I just fixed a bug in Bio::Align::DNAStatistics that failed with a div by zero error (bug 2901) on this data:
>> 
>>> seq1
>> GGTACCAGCAGGTGGGCCGCCTACTGCGCACGCGCGGGTTTGCGGGCAGCCGC
>>> seq2
>> GGTACCAGCTGGTGGGCCGCCTACTGCGCACGCGCGGGTTTGCGGGCAGCCGC
>>> seq3
>> GGTACCAGCAGGTGGTCCGCCTA------------------------------
>>> seq4
>> --------------------------CGCACGCGCGTGTTTGCGGGCAGCCGC
>> 
>> Since seq3 and seq4 don't overlap, the distance can't be calculated.  In our case, I replace the score with 'NA' as a placeholder, but I'm worried about downstream app breakage.  Anyone have an objection to using 'NA' here, or know of ways this may lead to problems elsewhere?
>> 
>> chris
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
> 
> --
> Jason Stajich
> jason.stajich at gmail.com
> jason at bioperl.org
> http://fungalgenomes.org/
> http://twitter.com/hyphaltip
> 
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l





More information about the Bioperl-l mailing list