[Genome] custom annotation error

Hiram Clawson hiram at soe.ucsc.edu
Wed Dec 19 10:32:28 PST 2007


Good Morning Mark:

The confusion you note is because of the 0-relative vs. 1-relative
coordinate systems in use at various stages of processing.

There are eight coordinates in your file that overlap with the
next specification, namely:
# chr10 98076015        98076029        -8.851
# chr10 98195233        98195247        8.261
# chr10 98301830        98301844        -8.874
# chr10 98332806        98332820        -9.174
# chr10 98496236        98496250        12.341
# chr10 98499874        98499888        8.213
# chr2  170149997       170150011       -10.221
# chr2  170163207       170163221       -9.469

When the error says, for example: previous: 98076028 > 98076017 <-current
it refers to the error when reading the next line after:
chr10 98076015 98076029 -8.851
which is:
chr10 98076017 98076031 8.781

Note how those two specifications overlap each other.
The 98076028 in the error message is from the chromEnd 98076029
in your input line, as is the chromStart 98076017 from the next line.
chromStart coordinates in a bed file are 0-relative, chromEnd
coordinates in a bed file are 1-relative.  The error message
is outputting everything in 0-relative coordinate space.
See also: http://genome.ucsc.edu/FAQ/FAQtracks#tracks1
and: http://genomewiki.ucsc.edu/index.php/Coordinate_Transforms

We can't draw two values in the same location, hence the overlap
error.  You will have to decide on which value you would like
to graph in the overlapping coordinate area.

I also note that all of your data points occupy 14 bases each.
Since all of your data points are the same size, you can
more efficiently submit this data with a variable step format,
for example, using the conversion procedure mentioned at:
http://genomewiki.ucsc.edu/index.php/Wiggle_BED_to_variableStep_format_conversion
you can convert your bed file into variable step format data, which
would begin to look like:
track type=wiggle_0 autoScale=off name=bs_test description="binding site test" c
olor=0,0,255 visibility=full
variableStep chrom=chr10 span=14
98046839        8.607
98047002        -9.89
98047906        11.702
... etc ...
variableStep chrom=chr2 span=14
170110809       -11
170111046       9.772
170111423       10.221
... etc ...
variableStep chrom=chrX span=14
158187001       11.598
158187689       10.929
158198757       -11.632
... etc ...

Your small example check.txt file becomes half the size
when converted.  This would become more important if your data input
is very large.  Plus, you can gzip compress these data inputs
to save network transfer time for very large data sets.

--Hiram

Mark Meyer wrote:
> I am attempting to create a .wig file format and have followed all criteria
> listed, however I am receiving this error when I am attempting to upload
> this file (attached)
> 
> *Error File ' check.txt' - track load error:
> chrom positions not in numerical order at line 21. previous: 98076028 >
> 98076017 <-current
> 
> *the position 98076028 does not even exist in this file.  All duplicates
> have been removed prior to upload.  It has done this with every file that I
> have tried to upload of my own data and every time the number it says is
> "not in numerical order", does not exist in the document.
> 
> Please help?
> 
> Thanks,
> 
> Mark


More information about the Genome mailing list