Programmatic access to the Genome Browser

The UCSC API for retrieving data and uploading data is RESTful over HTTP but does not use JSON to save computational time on the server. Download of most data formats requires client-side C tools that convert to/from binary files. Data upload uses custom text files.

Here are some common tasks that can be done from scripts with the UCSC Genome Browser. It is assumed that the reader knows the standard Unix command line tools.

Download data stored in a database table

• use Tools - Table Browser - "Describe schema" to browse the database schema. All fields have a human readable description and the links to other tables are shown.
• to access the public Mysql server, use a commen like mysql hg19 --no-defaults -h genome-mysql.cse.ucsc.edu -u genome -A -e 'select * from pubsBingBlat' -NB > out.txt
• the list of data tracks is part of the table trackDb. The first column is the internal name of the track.
• tracks with types that start with "big" are stored in files, all others are stored at least to some extent in Mysql tables.
• to use the genomic coordinates of non-big tables, retrieve the table and use our documentation of file formats http://genome.ucsc.edu/FAQ/FAQformat.html
• the first column in many tables with genomic coordinates is called "bin" and can be stripped for most applications

 mysql --no-defaults -h genome-mysql.cse.ucsc.edu -u genome -A -e "select ta
bleName, type, priority from trackDb where tableName in ('gold', 'refGene','knownGene', 'ccds', 'clinvar') limit 5"  hg19 
+-----------+-------------------------------------+----------+
| tableName | type                                | priority |
+-----------+-------------------------------------+----------+
| clinvar   | bigBed 12 .                         |      100 |
| gold      | bed 3 +                             |      100 |
| knownGene | genePred knownGenePep knownGeneMrna |        1 |
| refGene   | genePred refPep refMrna             |        2 |
+-----------+-------------------------------------+----------+

mysql --no-defaults -h genome-mysql.cse.ucsc.edu -u genome -A -e "select * from knownGene limit 3"  hg19 
+------------+-------+--------+---------+-------+----------+--------+-----------+--------------------+--------------------+-----------+------------+
| name       | chrom | strand | txStart | txEnd | cdsStart | cdsEnd | exonCount | exonStarts         | exonEnds           | proteinID | alignID    |
+------------+-------+--------+---------+-------+----------+--------+-----------+--------------------+--------------------+-----------+------------+
| uc001aaa.3 | chr1  | +      |   11873 | 14409 |    11873 |  11873 |         3 | 11873,12612,13220, | 12227,12721,14409, |           | uc001aaa.3 |
| uc010nxr.1 | chr1  | +      |   11873 | 14409 |    11873 |  11873 |         3 | 11873,12645,13220, | 12227,12697,14409, |           | uc010nxr.1 |
| uc010nxq.1 | chr1  | +      |   11873 | 14409 |    12189 |  13639 |         3 | 11873,12594,13402, | 12227,12721,14409, | B7ZGX9    | uc010nxq.1 |
+------------+-------+--------+---------+-------+----------+--------+-----------+--------------------+--------------------+-----------+------------+

Get the chromosome sequence for a range

• Download the tool twoBitToFa from http://hgdownload.cse.ucsc.edu/admin/exe/ e.g. with curl http://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/twoBitToFa > twoBitToFa; chmod a+x twoBitToFa. For OSX, please adapt the download location to http://hgdownload.cse.ucsc.edu/admin/exe/macOSX.x86_64/
• To get the DNA sequence from e.g. the human genome hg19, run a command like twoBitToFa http://hgdownload.cse.ucsc.edu/gbdb/hg19/hg19.2bit stdout -seq=chr21 -start=1 -end=10000. You can replace stdout with a filename of your choice.
• for best performance, download the 2bit file for your genome from http://hgdownload.cse.ucsc.edu/gbdb/ to local disk

Get the "wiggle" (x-y-plot) graph data for a chromosome range

• Download bigWigToWig from http://hgdownload.cse.ucsc.edu/admin/exe/ as shown above
• run a command like bigWigToWig http://hgdownload.cse.ucsc.edu/gbdb/hg19/bbi/wgEncodeBroadHistoneK562Cbx2Sig.bigWig -chrom=chr21 -start=0 -end=10000000 stdout. You can also replace stdout with a filename of your choice.

Get a copy of the current Genome Browser image from a script

• use curl http://genome.ucsc.edu/cgi-bin/hgRenderTracks > test.png. hgRenderTracks understands the same parameters and options as the main hgTracks CGI, e.g. <internalTrackName>=pack
• to get the internal track name of a track, mouse over the track and look at your internet browser status line or go to the track configuration page and look for the value of the variable called "g" in the current URL
• to hide the default track when you use hgRenderTracks, make sure that the first track parameter is hideTracks=1
• for example, to download the image for a chromosomal location with only the RefSeq transcripts and publications track to "pack" mode, use this command: curl 'http://genome.ucsc.edu/cgi-bin/hgRenderTracks?position=chr17:41570860-41650551&hideTracks=1&refGene=pack&pubs=pack' > temp.png

Upload a custom track and link to the genome browser with the track loaded

• create a custom track file as documented here http://genome.ucsc.edu/goldenpath/help/customTrack.html, e.g.: printf 'track name="TestTrack" description="TestTrack with links on features" url="http://www.google.com/$$"\nchr1 1 1000 testIdForUrl' > temp.bed
• upload your file with a command like this, it will print a string to stdout which we are calling $HGSID in the following curl -s -F db=hg19 -F 'hgct_customText=@temp.bed' http://genome.ucsc.edu/cgi-bin/hgCustom | grep -o 'hgsid=[0-9]*_[a-zA-Z0-9]*' | uniq | sed -e 's/hgsid=//'
• you can link to a fresh genome browser session with only this track loaded with http://genome.ucsc.edu/cgi-bin/hgTracks?hgsid=$HGSID&position=chr1:1-1000
• you can load more tracks into this session by adding the parameter hgsid=$HGSID to all future curl calls with the parameter "-F hgsid=$HGSID"
• you can download the image of this session with hgRenderTrack as shown above, by supplying the $HGSID value to hgRenderTracks, like this curl "http://genome.ucsc.edu/cgi-bin/hgRenderTracks?hgsid=$HGSID" > test.png