[ensembl-dev] Coordinates for imported BED track data

Sébastien Moretti sebastien.moretti at unil.ch
Wed Feb 2 14:59:51 GMT 2011


I have just tested BED tracks with ensembl 61 and it seems that the 
coordinates problem is still there.

I can use GFF to get right coordinate mapping but BED tracks allow nice 
features like color choice for each line.

>> Hi Sébastien
>> We are aware of this problem. Before we start modifying coordinates so
>> we can fix this we need to make sure that the automatic file format
>> detection always correctly identifies BED files. We are going to
>> modify the code to look at the file extension of the uploaded file as
>> well as the contents to determine format, and then for files
>> identified as BED files, adjust the coordinates internally from BED to
>> Ensembl convention. We hope to implement this for the next release.
>> Steve
> Thanks
> I will wait for the next release.
>>> Hi,
>>> I don't know if BED tracks are officially supported in Ensembl.
>>> It seems to be the case because such data can be imported.
>>> Nevertheless, coordinates from BED tracks seem to be misinterpreted:
>>> For BED tracks, the first base in a chromosome is numbered 0.
>>> For GFF, the first base in a chromosome is numbered 1.
>>> http://genome.ucsc.edu/FAQ/FAQformat#format1
>>> http://genome.ucsc.edu/goldenPath/help/customTrack.html#BED
>>> It means that this SNP, in GFF
>>> chr20 SNP synonymous_variation 35464269 35464269 ...
>>> must have these coordinates in BED
>>> chr20 35464268 35464269 synonymous ...
>>> In the image for this region, SNPs from BED tracks are drawn on 2
>>> nucleotides instead of one.
>>> It works if the BED track looks like this
>>> chr20 35464269 35464269 synonymous ...
>>> but it is no more an "official" BED format.
>>> (even if this fake BED syntax should continue to be interpreted for
>>> single position at least)
>>> How to solve this ?
>>> Thanks

Sébastien Moretti
Department of Ecology and Evolution,
Biophore, University of Lausanne,
CH-1015 Lausanne, Switzerland
Tel.: +41 (21) 692 4221/4079

More information about the Dev mailing list