Hi there, further to the query below, I'm trying to follow the steps outlined in the email below to send the 28way genome alignment to galaxy. I'm currently waiting on this job to complete on galaxy - as mentioned below, it does take some time, but can anybody give me an idea of how long this might take (its already been running for a while)? Is it likely to be minutes/hours/days? Cheers Dan
-------- Original Message -------- Subject: Re: [Genome] Chimp assembly 2 pairwise alignments Date: Mon, 07 May 2007 16:00:42 -0700 From: Archana Thakkapallayil email@example.com Reply-To: firstname.lastname@example.org To: Daniel Gaffney email@example.com CC: firstname.lastname@example.org References: 463F58BC.email@example.com
Here is the response from one of our engineers:
I think that the easiest way to do this is using Galaxy 2. This method will involve using the pairwise alignments extracted from the multiz multiple alignment (from the Conservation track). This is a best in genome alignment with the best aligning region for chimp aligned to each region of the human genome. This means that there is less information than in the chain track which contains all alignments (passing certain criteria) of the chimp genome to each region of the human genome. If this is ok, then you can use Galaxy by going to: http://test.g2.bx.psu.edu/
You will need to use the test version of Galaxy because this has the option of using our genome-test server. There are only panTro1 alignments in the 17-way alignment on hg18, but there are panTro2 alignments in the 28-way on hg18. Be aware that this our test server and that some of the data here (including the 28-way alignment) has not gone through our rigorous quality assurance checks.
Here are the instructions to follow:
1) Go to http://test.g2.bx.psu.edu/ 2) Click on "Get Data" in the pane at the left side. 3) Click on "UCSC Test table browser" link in the expanded list. 4) In the middle an interface will appear that looks like the Table Browser. You can then select the Human Mar 2006 (hg18) assembly. Select Comparative Genomics as the group, 28-way Cons as the track and "multiz28wayAnno" as the table. Make sure that genome is selected as the region. Then select MAF as the output type. 5) Check the box that says "Send output to Galaxy". Press "get output" button and the "Send query to Galaxy" button. (Up to this point can also be done in the Table Browser on genome-test). Getting the MAF output seems to be rather slow. 6) Then from the "Get Data" menu, select "Upload File" and you can paste in or upload a file of the regions for which you want alignments (position or BED format is fine). 7) Then click on the "Fetch Sequences and Alignments" in the left pane.
From the expanded list, select "Extract MAF blocks". First select the
genomic intervals (query 2) and then press "Next Step" and select the MAF format output (query 1). The result is the multiple alignments for the regions defined by the user. 8) To limit this to only the pairwise alignment between human and chimp, select the "Maf Limit to Species" from the left pane and choose hg18 and panTro2.
If you have questions about using Galaxy, you can contact their help desk: firstname.lastname@example.org
If you would rather use the human/chimp chain or net alignments to get alignments for specific regions of the human genome, then we can walk you through using several of the programs that we have available for command line use in order to extract human/chimp pairwise alignments from this data.
I hope this information is helpful to you. Let us know if you have further questions.
Archana UCSC genome Bioinformatics Group
Daniel Gaffney wrote:
Hi there, I have ~ 1500 small regions in the human genome, for which I would like to get the human-chimp pairwise alignment for the latest chimp assembly. However, I'm having a little trouble in figuring out the best way to do this - can this be done locally? Or is there a batch submission tool that I can retrieve all alignments from UCSC directly - I've tried the table browser, but I can't seem to get the specific region I want (i.e. if i select the chimp chain track, and retrieve sequence, I get the the alignment of that entire region, not just the region corresponding to the coordinates I've entered). Any ideas? Cheers Dan
I'm not sure I'm following the steps entirely. It looks like here you are trying to send the entirety of the 28-way alignments from the genome browser to Galaxy. This would require moving more than 30 gigs of data from the table browser to Galaxy, that would take a very long time.
Try the following steps instead:
1) Upload your intervals using Get Data -> Upload File. Ensure the file format of the new dataset is "bed" or "interval".
2) Go to Fetch Sequences and Alignments -> Extract MAF Blocks (*not* Extract User MAF Blocks). On the first page, select your dataset and click "Next Step". On the second page, choose "28-way multiz" and click Execute.
For 1500 intervals this shouldn't take more than a minute or so. Let me know how it goes.
On May 9, 2007, at 10:47 AM, Daniel Gaffney wrote:
- In the middle an interface will appear that looks like the Table
Browser. You can then select the Human Mar 2006 (hg18) assembly. Select Comparative Genomics as the group, 28-way Cons as the track and "multiz28wayAnno" as the table. Make sure that genome is selected as the region. Then select MAF as the output type. 5) Check the box that says "Send output to Galaxy". Press "get output" button and the "Send query to Galaxy" button. (Up to this point can also be done in the Table Browser on genome-test). Getting the MAF output seems to be rather slow.