On Tue, May 24, 2011 at 9:57 AM, Angie Hinrichs <[email protected]> wrote: > Hi Brent, > > Regarding this part of your original email: > >> chrom,txStart,txEnd,cdsStart,cdsEnd,K.name,X.geneSymbol,proteinID,strand,exonStarts,exonEnds >> from knownGene as K,kgXref as X where X.kgId=K.name > ... >> I have seen this: http://genome.ucsc.edu/FAQ/FAQtracks#tracks1 so the >> internal format matches the BED format, correct? > > Actually, those columns are from an internal format that we call genePred. > BED has a bit different column order, and it has a very different encoding of > block coordinates: instead of lists of starts and ends like genePred, BED has > a list of block sizes followed by a list of block starts -- but each block > start is relative to the overall chromStart (analogous to genePred's > txStart), so block starts always begin at 0. > > The Table Browser has BED as one of its output formats, so it's the easiest > way to get BED for a genePred table (or just about any non-BED table that has > chrom, start, end and other fields). > > Hope that helps, > Angie > >
Aye, I didn't ask that part clearly; just meant that it uses 0-based, half-open intervals like bed. Originally I was thinking maybe I was misunderstanding the format and that's why I was seeing the funky introns. _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
