Hi Alexander, So, you are expecting the output coordinates to cover only the bases where knownGene exons intersect with your items? You can do that.
When creating the intersection, select "Base-pair-wise intersection (AND) of User Track and UCSC Genes". Unfortunately, that loses the identity of the items and outputs only position ranges with auto-generated names (chr1.1, chr1.2, etc) as stated in the table browser: These combinations will discard the names and gene/alignment structure (if any) of "YourData" and produce a simple list of position ranges. With the default intersection options, we output the entire item from the primary track if it (or any of its exons, if it has exons) overlaps with any (exon of an) item in the secondary track. Please let us know if you have any additional questions: [email protected] - Greg Roe UCSC Genome Bioinformatics Group On 11/15/11 12:15 PM, Alexander Belostotsky wrote: > Thank you, Greg, > > I meant completely another thing that Table Browser seems to contain a > bug intersecting let say NOT ALL blocks by default when we think that > it must intersect all of them. > > Just take a look at this if you are interested. > > Alexander > > 2011/11/15 Greg Roe <[email protected] <mailto:[email protected]>> > > Hi Alexander, > > I'm not completely sure what you're trying to do. If you want to > intersect only cdsStart and cdsEnd (not txStart/txEnd, nor the > blocks) with some other table, then you can try using selected > fields to get only those coordinates, and then rearrange the > columns to get BED4 (or if you need BED6, rearrange and add a > dummy score column with 0's). Then upload that as a custom track > and do the intersection. > > If that's not what you're asking, Galaxy > (http://main.g2.bx.psu.edu/) or BEDtools > (http://code.google.com/p/bedtools/) might have a solution. > > > > Please let us know if you have any additional questions: > [email protected] <mailto:[email protected]> > > - > Greg Roe > UCSC Genome Bioinformatics Group > > > On 11/2/11 1:08 AM, Alexander Belostotsky wrote: >> Thank you very much! >> >> As far as I understand from your letter, Table Browser searches >> for intersection between all blocks by default. It is what I >> really need. But as a result I obtain intersection between >> pos_left and pos_right of some table and ONLY CDS of knowngene. I >> checked this out. >> >> So could you just tell me how can I see intersections between >> some table pos_left and pos_right and txStart and txEnd of >> knowngene by using Table Browser. >> >> Thank you! >> Alexander >> >> 2011/11/1 Greg Roe <[email protected] <mailto:[email protected]>> >> >> Hi Alexander, >> >> One of our engineers commented: >> " When a table has block structure (e.g., exons in genePred >> or bed, gapless aligned blocks in psl), the Table Browser >> intersects only the blocks instead of the entire span... and >> all blocks, not just the CDS portion. Unless the primary >> table is a wiggle... that does the entire span of the >> secondary table items regardless of block structure! Of >> course it would be nicer to offer the option of whole span or >> blocks (or CDS when known), but I don't know of any plans to >> add that functionality." >> >> Please let us know if you have any additional questions: >> [email protected] <mailto:[email protected]> >> >> - >> Greg Roe >> UCSC Genome Bioinformatics Group >> >> >> On 10/31/11 5:21 AM, Alexander Belostotsky wrote: >>> Hello, dear members of glorious UCSC Genome Browser team :) , >>> >>> I think it would be nice to report that *results of intersection >>> between >>> some table and "knowngene" table in Table Browser give results of >>> intersection between this table and CDS coordinates in "knowngene" >>> table.* >>> At least this occurs in hg18 assembly. >>> I suppose that people who use this want to have intersection >>> between some >>> table and TX coordinates. So, it should be intersection between >>> some let >>> say (pos_left between txStart and txEnd) and (pos_right between >>> txStart and >>> txEnd). >>> >>> Am I right. If you think yes, please settle the problem. >>> >>> Best wishes, >>> Alexander >>> _______________________________________________ >>> Genome maillist [email protected] <mailto:[email protected]> >>> https://lists.soe.ucsc.edu/mailman/listinfo/genome >> >> > _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
