[Genome] a bug?

[Ann Zweig]

Hello Silvia,

     This is an interesting question and I spent some time looking into it.  And 
now I have an explanation for you.

     When I used Blast on the NCBI site to Blast the sequence from BC088267 
against rat, I got a very good hit, as you did.  The reason for this is because 
it is Blasting against the entire genBank library of mRNAs.  It does align well 
to the mRNA, but I'm pretty sure that it's not really aligning to the actual rat 
genome.

     Now, when I use the UCSC browser to BLAT the sequence onto the rat assembly 
(rn4), it BLATs very poorly, as you noticed.  In fact this mRNA does not meet 
our requirements for alignment to the rat genome.  It does, however, meet our 
less-stringent requirements for aligning to the human genome, which are: "Only 
those alignments having a base identity level within 1% of the best and at least 
25% base identity with the genomic sequence".

     The reason that you cannot simply find a hit for BC088267 when you search 
for it in the rat assembly is because it does not meet minimum alignment 
criteria on this assembly.  If you search instead for 'distrobrevin', you will 
get a hit in the rat assembly, but the details page will explain that the mRNA 
did not align well enough to the assembly.

     If you view the Rat Chain track in the human assembly, at the location of 
your mRNA, chr6:15,373,603-15,991,950, you will see that it is in the middle of 
a large gap in a chain to the rat assembly on chr17.  So, there is a large chain 
from the human chr6 to the rat chr17, but this mRNA falls in a part of the human 
sequence that does not align well at all to rat.  If you click through from the 
top (pink) section of the Rat Chain track to the details page, then click on 
"View details of parts of chain within browser window." you see the gaps in the 
rat assembly if you turn on the Gap track.  The conclusion is, that the rat 
assembly does not yet contain the DNA to match this mRNA.

     I hope this helps explain the anomalies you noticed.  Please be sure to 
write back if you have more questions.

Regards,

----------
Ann Zweig
UCSC Genome Bioinformatics Group
http://genome.ucsc.edu


Silvia Jimenez wrote:
> Dear Sir
> 
> Is there a problem with 'BLAT'?
> 
> 1) I have the following GenBank/EMBL sequence BC088267;  and I wanted to 
> 'Blat' it but I retrieve no results (very bad matches on the genome). On 
> the other hand if I take this sequence and perform a blast on the Rat 
> genome at the NCBI the sequence perfectly matches the genome.(this 
> happened only with Rat Blat).
> 
> 2) There was another problem that seems to be resolved, after "blating" 
> when one went to details and when one wanted to click on together, there 
> was an error but it seems to be OK now.
> 
> Many thanks for your useful tool, there is a life before and after Blat.
> 
> Silvia
>