Re: [bedtools-discuss] weighted average of unionBedGraphs is desirable

[Aaron Quinlan]

Hi Sol,

This sounds interesting.  In order to increase my understanding of what you are looking for, would you mind providing an example of the output you to be produced by these options?

Best,

- Aaron

quinlanlab.org

On Apr 26, 2013, at 1:45 PM, solka...@gmail.com wrote:

I posted this as a comment to on the docs page:

  http://bedtools.readthedocs.org/en/latest/content/tools/unionbedg.html

A really desirable additional feature for unionBedGraphs would be to create a weighted average column.

Syntax could be "-weights wgt1 wgt2..."

And as an alternative, allow reciprocal weights, call them sizeFactors (as produced by DESeq) "-sizes size1 size2..."

An option "-avgOnly" would save the step of selecting only the average column.

The motivation for this is to create a track that is representative of the replicates of one condition, normalized among the replicates (by millions of mapped reads, or DESeq sizeFactors, or whatever).

Since this requires a floating point operation, "-digits" would give the desired number of decimal places (default 0, i.e. integer results).

Note that the weights need not sum to 1. Because you might want to compare the (weighted) average of one condition to the (weighted) average of another condition, where there is one set of weights for all samples across multiple conditions.

/Sol Katzman

--
You received this message because you are subscribed to the Google Groups "bedtools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bedtools-discu...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.