[Genome] Blat finds NO match for genomic DNA

[Rachel Harte]

Hi Ulas Karaoz,

No, you are not doing anything wrong. The reason this is happening is to 
do with the settings used for the Blat tool that you use through the web 
interface. If you look at this region that you are using for Blat in the 
Genome Browser, you will notice that it is entirely composed of repeats.
When using Blat through the web interface, there is a setting that is used 
that involves repeats and that is the repMatch option that is set by 
default to 1024 for a tile size of 11 which is the default tile size. The 
tile size of 11 bp is the size of the the match that triggers an 
alignment. repMatch sets the number of repetitions of a tile allowed before
it is marked as overused and therefore not used to start an alignment.

This region that you happened to find has a large number tiles that 
exceeded the default repMatch=1024 limit and therefore an alignment 
candidate could not be triggered.

For more information about Blat, please see:
http://genome.ucsc.edu/FAQ/FAQblat

or you may read the Blat paper:
http://www.genome.org/cgi/content/abstract/12/4/656

I hope that this answers your question. Please let us know if you have 
further questions.

Rachel

On Tue, 24 Oct 2006, Ulas Karaoz 
wrote:

> Hi,
> I have the following observation, which is unexpected. I pull out the
> following human DNA using the browsers "Get DNA in a window"
> (range=chr17:34993079-34993128 in hg17-May2004) and then Blat the same
> sequence against the same version of the genome. Blat cannot find any
> match.
>
> How is that possible? Am I doing anything wrong?
>
> The sequence I am pulling out via
> "http://genome.ucsc.edu/cgi-bin/hgc?hgsid=79481900&o=34993078&g=getDna&i=mixed&c=chr17&l=34993078&r=34993148&db=hg17&hgsid=79481900"
> is as follows:
>
>> hg17_dna range=chr17:34993079-34993128 5'pad=0 3'pad=0 revComp=FALSE strand=? repeatMasking=none
> ACCCCATCTCTACTAAAAATACAAAAATTAGCTGGGTGTGGTGGCCCATG
>
>
>
>
> Best.
> Ulas Karaoz.
>
> _______________________________________________
> Genome maillist  -  Genome at soe.ucsc.edu
> http://www.soe.ucsc.edu/mailman/listinfo/genome
>

-- 
Rachel Harte
UCSC Genome Bioinformatics Group
http://genome.ucsc.edu