That works! I used GTF output instead, because that's what I'm trying to feed into my current script, but that doesn't seem to alter the overall content of your approach.
Thanks, Steve! I'll follow up on the other thread to map back. On Mon, Feb 6, 2012 at 3:24 PM, Steve Heitner <[email protected]> wrote: > Hello, John. > > You can do this by using our Table Browser. If you're unfamiliar with the > Table Browser, please see the User's Guide at > http://genome.ucsc.edu/goldenPath/help/hgTablesHelp.html. > > It sounds like what you're trying to do is to exclude entries where the > entry for "transcript" in the "knownCanonical" table is "n/a". You'll need > to create a filter to exclude the results than have no knownCanonical > transcripts. Follow the below steps: > > 1. From http://genome.ucsc.edu, select "Tables" from the blue navigation > bar > at the top of the screen. > > 2. Select the following options: > Clade: Mammal > Genome: Human > Assembly: Feb. 2009 (GRCh37/hg19) > Group: Genes and Gene Prediction Tracks > Track: UCSC Genes > Table: knownGene > Region: Select either "genome" for the entire genome or specify a region > next to "position" > Output format: "selected fields from primary and related tables" > > 3. Click "create" on the "filter" line > > 4. Scroll down to the "Linked Tables" section and check the > hg19.knownCanonical checkbox > > 5. Scroll to the bottom of the page and click the "Allow Filtering Using > Fields in Checked Tables" button > > 6. In the "hg19.knownCanonical based filters" section, set the following on > the transcript line: transcript doesn't match n/a > > 7. Click the "submit" button > > 8. Click the "get output" button > > 9. Select the fields you would like to include in your results. Selecting > the hg19.kgXref.geneSymbol field will display the proper gene name in your > results. > > 10. Click the "get output" button > > Please contact us again at [email protected] if you have any further > questions. > > --- > Steve Heitner > UCSC Genome Bioinformatics Group > > -----Original Message----- > From: [email protected] [mailto:[email protected]] On > Behalf Of John Hawkins > Sent: Wednesday, February 01, 2012 3:13 PM > To: [email protected] > Subject: Re: [Genome] Question about knownCanonical / refFlat / GTF format > > I'm trying (and failing) to get the UCSC genome browser to output what I > need, and would appreciate guidance. > > My goal is to get essentially the refFlat table data for the knownCanonical > transcripts. > > In other words, I want to get GTF encoded information about transcripts, > with the gene/transcript names flattened into the data, but I only want to > see data for canonical transcripts, not all possible. > > Can this be done directly? If not, any suggestions for how to most simply > approximate it by merging things on my end? > > The settings I've tried look like this: > > > > but this gives me lines like: > > chr1>---hg19_knownCanonical>----exon>---322037>-326938>-0.000000>------- > chr1>.>------.>------gene_id > "gene9"; transcript_id "tx9";- > > where "gene9" and "tx9" are just iterative line-by-line counters that are > clearly meaningless because they never repeat. Meaning that each exon has a > unique (and numerically matched) "transcript" and "gene" label with no > semantic content. > > And also I don't appear to get strand information. > > In the past I know I've been able to get BED files that contained strand > info and at least grouped exons by transcript, though I don't recall the > settings that got me there... > > Anybody see what I r doin rong, here? > > Thanks, > > -John > _______________________________________________ > Genome maillist - [email protected] > https://lists.soe.ucsc.edu/mailman/listinfo/genome > > -- "mFphrrrbm hoAmn rggh, hnllmmgh rnbnrrl." --Demosthenes, "The Early Works" _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
