[Genome] difference of local Blat and web-Blat

[Archana Thakkapallayil]

Hello Xianjun,

One of our developers suggested the following:

1. It is NOT recommended to use BLAT across such a large evolutionary 
distance (human micro-rna to zebrafish genome).  Even if one were to 
attempt large distances, protein BLAT would be more sensitive -- 
although you are doing RNA, so it may have different constraints.

blastz from Webb Miller at Pennsylvania State University is what we use 
for cross-species alignments, especially for distances further apart 
than human-mouse.

2. If you are going to use blast8 output, you don't have a .psl and 
shouldn't use that extenison. Also blast8 is not going to give exactly 
the same output as psl.  For example, psl is a linked-features format 
where the exons have been chained together.  Blast output is single-exon.

3. Using the following command, I was able to exactly reproduce the same 
results that hgBlat gives on our site for danRer5 BLAT :

blat /gbdb/danRer5/danRer5.2bit test.fa testdr5.psl -stepSize=5 
-minScore=0 -minIdentity=0

This shows that command-line blat and hgBlat are working.

I hope this information is helpful to you. Please don't hesitate to 
contact us again if you require further assistance.

Regards,

Archana
UCSC Genome Bioinformatics Group


>
> Subject:
> difference of local Blat and web-Blat
> From:
> Xianjun Dong <Xianjun.Dong at bccs.uib.no>
> Date:
> Tue, 20 Nov 2007 16:53:37 +0100
> To:
> genome at soe.ucsc.edu
>
> To:
> genome at soe.ucsc.edu
>
>
> Dear sir./madam,
>
> I am trying to use local Blat for my miRNA(~70bp in length) sequences, 
> with the following parameters
>
> blat /export/data/goldenpath/danRer5/assembly.2bit test.fa testdr5.psl 
> -out=blast8 -stepSize=5 -minScore=0 -minIdentity=0
>
> as the FAQ page (http://www.genome.ucsc.edu/FAQ/FAQblat.html#blat5) said.
>
> But the output is still different from that of the web-based Blat, 
> which are
>
> ================= local Blat =================
> hsa-mir-137     chr24   96.43   84      3       0       9       
> 92      35690482        35690399        3.4e-37 153.0
> hsa-mir-137     chr17   100.00  16      0       0       59      
> 74      14974006        14973991        2.1e+00 30.0
>
> ================= web-based Blat =============
>    ACTIONS      QUERY           SCORE START  END QSIZE IDENTITY CHRO STRAND  START    END      SPAN
> ---------------------------------------------------------------------------------------------------
> browser <http://www.genome.ucsc.edu/cgi-bin/hgTracks?position=chr24:29270667-29270750&db=danRer5&ss=../trash/hgSs/hgSs_www_7e6c_301a00.pslx+../trash/hgSs/hgSs_www_7e6c_301a00.fa&hgsid=99930361> details <http://www.genome.ucsc.edu/cgi-bin/hgc?o=29270666&g=htcUserAli&i=../trash/hgSs/hgSs_www_7e6c_301a00.pslx+..%2Ftrash%2FhgSs%2FhgSs_www_7e6c_301a00.fa+hsa-mir-137&c=chr24&l=29270666&r=29270750&db=danRer5&hgsid=99930361> hsa-mir-137       78     9    92   102  96.5%    24   -   29270667  29270750     84
> browser <http://www.genome.ucsc.edu/cgi-bin/hgTracks?position=chr24:28811094-28811177&db=danRer5&ss=../trash/hgSs/hgSs_www_7e6c_301a00.pslx+../trash/hgSs/hgSs_www_7e6c_301a00.fa&hgsid=99930361> details <http://www.genome.ucsc.edu/cgi-bin/hgc?o=28811093&g=htcUserAli&i=../trash/hgSs/hgSs_www_7e6c_301a00.pslx+..%2Ftrash%2FhgSs%2FhgSs_www_7e6c_301a00.fa+hsa-mir-137&c=chr24&l=28811093&r=28811177&db=danRer5&hgsid=99930361> hsa-mir-137       78     9    92   102  96.5%    24   +   28811094  28811177     84
> browser <http://www.genome.ucsc.edu/cgi-bin/hgTracks?position=chr2:15780390-15780478&db=danRer5&ss=../trash/hgSs/hgSs_www_7e6c_301a00.pslx+../trash/hgSs/hgSs_www_7e6c_301a00.fa&hgsid=99930361> details <http://www.genome.ucsc.edu/cgi-bin/hgc?o=15780389&g=htcUserAli&i=../trash/hgSs/hgSs_www_7e6c_301a00.pslx+..%2Ftrash%2FhgSs%2FhgSs_www_7e6c_301a00.fa+hsa-mir-137&c=chr2&l=15780389&r=15780478&db=danRer5&hgsid=99930361> hsa-mir-137       78     6    92   102  95.5%     2   +   15780390  15780478     89
>
> You see the result are much different, where the web-based one is more reasonable. 
>
> My query sequence is fairly simple:
> >hsa-mir-137 MI0000454
> GGTCCTCTGACTCTCTTCGGTGACGGGTATTCTTGGGTGGATAATACGGATTACGTTGTTATTGCTTAAGAATACGCGTAGTCGAGGAGAGTACCAGCGGCA
>
> The local Blat version is BLAT v. 34.
>
> Could you help to have a look? Thanks very much.
>
> Regards,
>
> Xianjun
>