[Bioperl-l] progress on Bio::Search::Tiling

[Mark A. Jensen]

Hi Sendu- thanks--
After I get some more tests together (and they pass), my plan is to move it over 
to core, barring any objections. I'll put up a HOWTO with some discussion and 
compare/contrast, per your suggestion.
cheers
----- Original Message ----- 
From: "Sendu Bala" <bix at sendu.me.uk>
To: "Mark A. Jensen" <maj at fortinbras.us>
Cc: "BioPerl List" <bioperl-l at lists.open-bio.org>
Sent: Wednesday, May 27, 2009 5:46 PM
Subject: Re: [Bioperl-l] progress on Bio::Search::Tiling


> Mark A. Jensen wrote:
>> With the current bioperl-dev revision of MapTiling (r15712), I get
>> 6/1095 "failures", which you may judge below. The results indicate to
>> me that the algorithm is working, and that the translated dna
>> coordinate transformations (which kick in based on study of the
>> algorithm name) are correct, at least in this limited set.
>>
>> Comments *highly* appreciated-- many thanks for any time you spend
>> considering this.
> [snip]
>> not ok 389 - tricky.wublast(1)
>> #   Failed test 'tricky.wublast(1)'
>> #   at blast-MapTiling.t line 571.
>> #     '0.222'
>> #         >
>> #     '0.9'
>> # Not sure why frac_identical should be > 0.9, when no reported 'Identities' 
>> value is greater than 24% /maj
>
> That's almost certainly my mistake. For those two cmp tests I just didn't know 
> what the answer was supposed to be, so probably picked a figure out of thin 
> air based on the subsequent 2 answers.
>
> You should manually (or by some other independent method) check what the 
> correct answers are for those two cmp tests and change them to 'is' tests.
>
> Thanks for your work on this. I'd be interested in seeing a summary of what 
> you've achieved at the end. Especially if you point out areas where the old 
> system was getting things wrong - people may want to recheck their results 
> with the new modules.
>
>