[GET-dev] Blosum100 matrix correct in GET?

[Kimberly Robasky]

Thanks so much Xiaodi, but oh so many new questions ---

Let's start with: why are we using the negative of the BLOSUM value?

-Kim

On Tue, May 25, 2010 at 2:04 PM, Xiaodi Wu <xiaodi.wu at gmail.com> wrote:
> I hadn't looked into it before. Keep in mind that we use the negative
> of the BLOSUM value; so we agree on G>A.
>
> As to the other ones, the discrepancy originates in the source data.
> The values are in substitution_matrix.py, and the BLOSUM100 there
> gives me G>V as -5 and G>R as -4. I got these values from the source
> posted at <http://portal.open-bio.org/pipermail/biopython/2000-September/000418.html>
> which is in turn derived, for BLOSUM100, from the no-longer
> functioning URL
> <http://www.embl-heidelberg.de/~vogt/matrices/blosum100.cmp>. I can't
> explain why that source gives those values for BLOSUM100. I guess it's
> just an inaccurate source, and we can modify it as necessary.
>
>
> On Tue, May 25, 2010 at 9:10 AM, Alexander Wait Zaranek
> <awaitz at post.harvard.edu> wrote:
>> I seem to remember you had looked into this.
>>
>>
>> ---------- Forwarded message ----------
>> From: Kimberly Robasky <krobasky at gmail.com>
>> Date: Tue, May 25, 2010 at 10:01 AM
>> Subject: [GET-dev] Blosum100 matrix correct in GET?
>> To: get-dev at lists.freelogy.org
>>
>>
>> I sent this to Tom & Sasha initially, and Xaodi is looking into it,
>> but Sasha encouraged me to post it here (where Xaodi can post a response):
>>
>> G->A, G->V, G->R in Blosum100 from NCBI are, respectively, -1, -8, -6:
>> ftp://ftp.ncbi.nih.gov/blast/matrices/BLOSUM100
>>
>> But on GET, its 1, 5, 4:
>> http://evidence.personalgenomes.org/RHO-G51A
>> http://evidence.personalgenomes.org/RHO-G51V
>> http://evidence.personalgenomes.org/RHO-G51R
>>
>> What am I missing?
>> -Kim
>>
>> _______________________________________________
>> GET-dev mailing list
>> GET-dev at lists.freelogy.org
>> http://lists.freelogy.org/mailman/listinfo/get-dev
>>
>
> _______________________________________________
> GET-dev mailing list
> GET-dev at lists.freelogy.org
> http://lists.freelogy.org/mailman/listinfo/get-dev
>