[Genome] Genome graph window size

[Galt Barber]

Genome Graphs

If you convert your BED file with a script to the
native Genome Graphs custom file format, which is
very simple, e.g.
chr pos value
chr12 120000 1

You can have up to millions of points, although
it was designed expecting more like hundreds of thousands.

If you only plot the peaks and they are all the
same value, you won't see them, but by adding a
few points around your peak you can give it definition.
i.e.
chr12 120000 1
chr12 130000 1
chr12 140000 1
chr12 150000 1
Since it just connects the dots,
definining only the peaks doesn't work.
It just joins them into a what looks like a single line.
(There is also a user-configurable limit value
that causes it to not connect two dots further apart
than the max #bp you specify.)

However, if you had features that are
say 100 bp wide and non-overlapping,
you can define each peaks with multiple points like this:
chr12 120000 0
chr12 120001 1
chr12 120099 1
chr12 120100 0
And now you have a square peak 100bp wide.

If you prefer a triangular peak, you could do this:
chr12 120000 0
chr12 120050 1
chr12 120100 0

Note that the genome graphs display has a limited
number of pixels so that you are probably looking
at roughly thousands of bp per pixel.

But when you click on it to go through to the
genome-browser, you will be able to zoom in to
any level of detail.

-Galt


On Thu, 10 Jan 2008, Hiram Clawson wrote:

> Good Morning Robert:
>
> If your data can be calculated to a consistent window size,
> say one data point for every 100bp, you can use the
> wiggle format to view the data in the genome browser:
> http://genome.ucsc.edu/goldenPath/help/wiggle.html
>
> The genome graphs are whole genome viewpoints and wouldn't
> be able to show your data detail below a couple 100 thousand
> bases.  Unless you view them in the genome browser, where
> you could instead view them as wiggle data plots.
>
> Wiggle data points are bar graphs at the specified
> locations with the bars of size "span".  (span must
> be consistent)  Whereas, the genome graph data
> points are points on a line graph.  They are different
> types of graphs.  The wiggle data can be fudged into
> looking like a line graph by setting the graph
> options to "points" (actually the lines of the tops
> of the bars) and set smoothing on to smear those little
> line segments into a continuous looking line.
>
> --Hiram
>
> Robert Darnell wrote:
> > Hi:
> >
> > I'm a molecular biologist (so have mercy please) interested in trying
> > to graph BED files uploaded into Genome Graphs, but to have the data
> > plotted in a narrower window than 10,000bp; even 1,000 bp would be an
> > improvement I think for our data, or, optimally, 100bp.  Is this
> > impossible because of computation time, or feasible?
> >
> > Thanks,
> >
> > Bob
> > _________________________________
> > Robert B. Darnell
> > The Rockefeller University
> >
> >
> > _______________________________________________
> > Genome maillist  -  Genome at soe.ucsc.edu
> > http://www.soe.ucsc.edu/mailman/listinfo/genome
> >
>
> _______________________________________________
> Genome maillist  -  Genome at soe.ucsc.edu
> http://www.soe.ucsc.edu/mailman/listinfo/genome
>