[Genome] Liftover dies when using BED files with 12 fields?

[Hiram Clawson]

Good Morning John:

Are you using the WEB site for this liftOver operation, or are
you using the liftOver binary in a local situation there ?

We're missing some functionality in some of our utilities that do
not properly recognize this column nine in its new form of r,g,b
numbers.  It used to only be a 0 at all times.  I have recently fixed
one of these situations, I would like to fix this one you mention.

As a work-around, although you lose your r,g,b specification, turn
that column 9 into a 0 and then it will lift.

--Hiram

John Major wrote:
> Hello UCSC folks-
> 
> I am trying to convert a large # of BED files from DM2 to DM3.
> These files are all BED files using all of the 12 allowed fields ( 
> http://genome.ucsc.edu/goldenPath/help/customTrack.html#BED ).
> 
> And example line from my BED file is:
> (space separated)
> chr4 646375 646462 miR-1000 0 + 646375 646462 255,0,0 2 42,41 0,46
> chrX 12896008 12896104 miR-1026 0 + 12896008 12896104 255,0,0 2 47,44 0,52
> 
> When I try to run liftover on these files, I get the following error:
> 
> --Reading liftover chains
> --Mapping coordinates
> --invalid unsigned number: "255,0,0"
> 
> According to the documentation for liftover, I should be able to use 
> valid BED files.  These BED files are displayed properly via the Genome 
> Browser, so am i missing something obvious here?  Do I need to enclose 
> the last 4 fields in some kind of quotes?  Or does liftover not actually 
> work with the last 4 BED file fields?
> 
> thanks!
> john major