[Genome] how is knowntoEnsembl generated?

[Hiram Clawson]

Good Morning Jean:

You can always find the sequence of building any table in the genome
browser in our doc files in the source tree:
	src/hg/makeDb/doc/*.txt

The knownToEnsembl is created via the program hgMapToGene:

hgMapToGene - Map a track to a genePred track.
usage:
    hgMapToGene database track geneTrack newMapTable
where:
    database - is a browser database (ex. hg16)
    track - the track to map to the genePred track
            This must be bed12 psl or genePred itself
    geneTrack - a genePred format track
    newMapTable - a new table to create with two columns
            <geneTrackId><trackId>
For each gene in geneTrack this will select the most
overlapping item in track to associate

And for the knownToEnsemble, typical usage is:
	hgMapToGene hg18 ensGene knownGene knownToEnsembl

--Hiram

Jean Chang wrote:
> Hi,
> 
> I'm not able to find anywhere how knowntoEnsembl is generated...
> 
> I found this:
>> The hg18 (March 2006/NCBI Build 36.1) knownToEnsembl table was  
>> built in May, 2006 using data downloaded from Ensembl BioMart at  
>> that time.  The hg17 (May 2004/NCBI Build 35.1) knownToEnsembl was  
>> built in May, 2005. We did not log specific Ensembl versions.
> 
> But what I'd really like to know is how you create the mapping of  
> genes using the Ensembl BioMart - or if you could point me to how  
> Ensembl says they do it. I took a look at the Biomart and couldn't  
> figure out where you would get those mappings but what I'm looking  
> for is the methods to those mappings.
> 
> Thanks,
> 
> Jean