================
Anyuan Guo Ph.D.
Postdoc Fellow
Virginia Institute for Psychiatric and Behavioral Genetics
Virginia Commonwealth University
P.O. Box 980126
Richmond, VA 23298-0126, USA
Email: aguo@vcu.edu

Hi Anyuan,

I don't believe it is possible to retain the RefSeq ID in this case when using the Table Browser.  However, I think that Galaxy has this capacity, either by doing the intersection from scratch using their tools, or by joining your MAF with your custom track based on the genome coordinates.

Galaxy has screencasts:
http://galaxy.psu.edu/screencasts.html

and a wiki:
http://g2.trac.bx.psu.edu/

This screencast might be particularly helpful:
http://screencast.g2.bx.psu.edu/galaxy/MAF_manipulation/

If you have more questions about how to accomplish your task using Galaxy, you can contact them at galaxy-user@bx.psu.edu.

Good luck with your research.

--
Brooke Rhead
UCSC Genome Bioinformatics Group


a lot of  at galaxy-user@bx.psu.edu for help.

On 12/10/08 15:17, Anyuan Guo wrote:

Dear Brooke,
   Thanks very much. I learned a lot about creating custom track in your email. I can download a ~76Mb compressed file when I follow your instruction to create a custom track for upstream 1000 bp of RefseqGene and intersect with 17-way Cons. But I found the file format is not begin with Refseq ID (NM_xxxx). The following is the first 4 lines of the file.
##maf version=1
a score=-55252.000000
s hg18.chr1                   14754  99 + 247249719 CTGTGGGTCGGAGCCGGAGCGTCAGAGC---------CACCCACGACCACCGGCACGCC----CCCACCACA-GGGCAGCGTGG-TGTTGAGACAAC------A

     In fact, I need a file begin with Refseq ID, the downloaded maf file (http://hgdownload.cse.ucsc.edu/goldenPath/hg18/multiz17way/upstream1000.maf.gz) exactly match my requirement. But because some refseq sequences were updated, the downloaded file is out of date.
   The following is the first 4 lines of the downloaded file, which I need.
##maf version=1 scoring=zero
a score=0.000000
s NM_198943 0 1000 + 1000 GCATTTTAAACCCAAGTG----AAATCTCCTAGG----------CCCTTCATGCCACACTCA-----TCCATCCCTACCTAC--TTGTGTTGCAACCAAGGGCCCCAC

How can I get the up-to-date version of this download file?
Thanks.

Anyuan


Brooke Rhead wrote:

Hello Anyuan,

The reason that the sequence is different via the download file and the Table Browser is that the sequence associated with NM_014223 at RefSeq has changed since the download file was made.  The items in the RefSeq Genes track are updated daily; the download files are generally only made once.

You can see the revision history for any GenBank accession at NCBI:
http://www.ncbi.nlm.nih.gov/entrez/sutils/girevhist.cgi?val=NM_014223

The download file was last updated on 7-7-2007.  I tried blatting the NM_014223 sequence from the "Jun 3 2007 1:10 PM" update to the hg18 assembly, and the sequence aligned starting at the genomic coordinate chr1:40,929,952.  The upstream sequence from the file you downloaded corresponds to the 1,000 bases upstream of that base.

You can get an up-to-date version of the download file by creating yourself with the Table Browser.  First, make a custom track of the upstream regions of RefSeq Genes.  If you select the RefSeq Genes track in the Table Browser and choose "output format: custom track", you will be presented with an option to create one BED record per region that is "Upstream by ___ bases".  Enter 1,000 or 2,000 in this box and hit "get custom track in genome browser".  You should see a new custom track containing blocks representing regions upstream of all RefSeq Genes.

Now you can intersect your new custom track with the multiz alignment in the Conservation track to get only the upstream regions.  To do this step, select the 17-way (or 28-way) Conservation track in the Table Browser.  Select the table 'multiz17way' and region: genome.  Hit the "intersection: create" button and select your custom track.  Choose the option for "Base-pair-wise intersection (AND) of 17-Way Cons and upstream regions from refGene" and hit submit.  Back on the main Table Browser page, select "output format: MAF".  The size of the file you will be creating is quite large (76 Mb compressed for 1,000 base regions).  I suggest entering a name for the file and selecting the option to get a gzip compressed version of it.  Hit "get output".  You should end up with a MAF file that contains only the regions upstream of RefSeq Genes.

You may also be interested in the tools for working with MAF alignments at Galaxy: http://galaxy.psu.edu/ .  Galaxy is run by our collaborators at Penn State and extends the functionality of the Table Browser.  For instance, there is a tool to filter any undesired species from a MAF file, leaving only the species of interest to you.

I hope this is helpful.  If you have further questions, please feel free to contact us again at genome@soe.ucsc.edu.  If you have questions specific to Galaxy, their helpdesk email address is galaxy-user@bx.psu.edu.

-- 
Brooke Rhead
UCSC Genome Bioinformatics Group



Subject: question or bug about UCSC genome browser sequence
From: Anyuan Guo <aguo@vcu.edu>
Date: Mon, 17 Nov 2008 10:54:21 -0800
To: genome@soe.ucsc.edu


Dear author,
    Thanks for you providing the wonderful database and website of UCSC
genome browser.
    I have question about the sequence in it.
       I downloaded the human upstream 1000bp multiz alignment file from
ftp://hgdownload.cse.ucsc.edu/goldenPath/hg18/multiz17way/upstream1000.maf.gz
    When I check my sequence id  NM_014223.
    I can find the upstream 1000 bp sequence of this refseq gene in the
downloaded multiz alignment file.
    I also can search this id in genome browser and get the upstream
1000 bp using the "DNA" or "Tables" menu at the top of genome browser page.
    But I find these two upstream 1000 bp sequence are totally
different. I think the one using genome browser is right.
    But I am not just need the upstream 1000bp sequence, I need the
alignment with mouse sequence.

    Can I just get the sequence alignment between human and mouse for
all the refseq gene and the upstream 1000 or 2000 of these genes? Where
can I find it?
    I think those ortholog gene alignment (including upstream
regulatory sequence alignment) between two popular genome will be very
useful.

thanks.

Anyuan