Generating UCSC screen dumps in batch

[frozenlyse]

http://onetipperday.blogspot.com.au/2012/07/get-ucsc-images-for-list-of-regions-in.html

[jiaco]

Yeah, something fun to do first thing this morning. Despite the functional hack by R Finney, here is another that gets you the image in pdf format.


Put the first code block in wgetUcscPdfs.sh and chmod +x it

Be sure to change the hgsid to a value from your browser session

#!/bin/bash

baseurl="http://genome.ucsc.edu/"
hgsid=12345678

cat $1 | while read chr beg end;
do
	position=$chr":"$beg"-"$end;
	pdf_fname="auto_"$chr"_"$beg"_"$end".pdf";

	s=$baseurl"cgi-bin/hgTracks?hgt.psOutput=on&hgsid="$hgsid"&position="$position
	wget $s -O XXX
	s2=`grep 'hgt_genome.*.pdf' XXX | sed 's/.*hgt_genome/hgt_genome/' | sed 's/[.]pdf.*/.pdf/'`
	tgt=$baseurl"/trash/hgt/"$s2
	wget $tgt -O $pdf_fname
done

Here is a sample input, albeit a stupid one, the regions are all Ns

chr1	50000	50100
chr22	100000	100100

And the command line to make it happen.

./wget_UcscPdfs.sh test_wgetUcscPdfs.txt

You can make the file names how you want and even clean up url escaping if you like. It uses 2 rounds of wget but in the end you can get pdfs of each region.

[Richard Finney]

(I tought it would be hard in that I thought the individual tracks had separate images)
Set up your session by hand. Set up the tracks you want and the track mode (squish, full, dense, whatever). Set up your custom track.
Cut the URL for an initial query to hgTracks (the genome you want and any old position).
You'll need an "hgsid" variable. Make sure you have this in your URL. Your command line wget will mimick
the user you are in a browser by using this hgsid variable.
Use that as base URL to hack. Hack position parameter as needed ...

Here's an example, you can parameterize the position using your favorite shell code ...

wget "http://genome.ucsc.edu/cgi-bin/hgTracks?hgHubConnect.destUrl=..%2Fcgi-bin%2FhgTracks&clade=mammal&org=Human&db=hg19&position=chr21%3A33%2C031%2C597-33%2C041%2C570&hgt.positionInput=chr21%3A33%2C031%2C597-33%2C041%2C570&hgt.suggestTrack=knownGene&Submit=submit&hgsid=317627943&pix=1180" -O x
grep png x | grep trash | grep png | sed 's/src=...//' | tr -d "'" | grep png | awk 'BEGIN{p=0};{print "wget \"http://genome.ucsc.edu/" $3"\" -O "p"\.png"; p++}' | grep 3.png > tmp.get.job
. ./tmp.get.job
cp 3.png final.png

# I thought the images were in chunks , but they are not, third image will do.
#the following code using imagemagick are not necessary to "concatenate images".
#odds=`ls 3.png 5.png 7.png 9.png 11.png 13.png 15.png 17.png 19.png 21.png 22.png 23.png 25.png 27.png`
#evens=`ls 2.png 4.png 6.png 8.png 10.png 12.png 14.png 16.png 18.png 20.png 22.png 24.png 26.png`
#convert $odds -append left.png
#convert $evens -append right.png
#convert left.png right.png +append final.png

This just wgets the URL to a file called "x" and gets the third (4th ?) png in the resulting file which you can then wget. The final.png is the file you want to look at. You can rename it as necessary, perhaps using a parameter to your script.

There are likely more elegant ways, but what the heck it's a hack that works.

Regardless this should be a good enough hint for some wizard to fix up.

Resulting image may need alpha channel removed. I think. I"m not sure. It looks transparent in gimp but appears okay in browser.
Use imagemagick for this .

[larsgt]

Generating UCSC screen dumps in batch Reply to Thread

You could script almost anything. Can you be more specific what is in your "batch"?

Sorry for being a bit unclear.

What I want to do is write a script that takes in a number of regions say
chr1:234234234-2342424234 and gives me back a PNG/JPG/whatever of
the UCSC browser with a few tracks and a custom track.

/Lars

[rfilbert]

You could script almost anything. Can you be more specific what is in your "batch"?

[larsgt]

Hello,

Is there a good way of dumping images from UCSC in batch?

Probably going to use some form of web testing tool but I was
wondering if there was an easier way.

/Lars