[Gmod-help] Re: [Gmod-gbrowse] coordinates in ucsc_genes2gff.pl output

[Mikhail Pachkov]

Hi Dave,

Sorry for late response. I have been away for a while.

> The comments say data from UCSC is 0 relative, it then adds 1 to compensate,
> and then, as I read the code it subtracts 1 again; thus canceling the
> adjustment (assuming you are using the default ORIGIN of 1).  Which does not
> make sense to me.  (Caveat: my perl is not stellar.)
> Have you tried running it with -origin 0 ?
> Dave C

According to what I have found in UCSC documentation, coordinates in
downloadable files (e.g. refGene.txt) are given in a special format:
all starts are 0-based coordinates, all ends are 1-based coordinates.
So adjusting stars by 1 at beginning of the script translate starts to
1-based coordinates. Now starts and ends are 1-based. After that
ORIGIN is subtracted from starts and ends. Why ORIGIN is subtracted I
do not understand. Anyway I set ORIGIN to 0 and get my features in
1-based coordinates. However in original code it is impossible to use
"-origin 0" option since it will be set 1 in this case. IMHO, it is a
bug. If you like I can send a patch which fixes that problem.

Best regards,

Mikhail