[BioC] LimmaGUI discards spots with no signal on one channel?

[Gordon K Smyth]

> Date: Fri, 08 Apr 2005 15:56:39 +0100
> From: J.delasHeras at ed.ac.uk
> Subject: [BioC] LimmaGUI discards spots with no signal on one
> 	channel??
> To: bioconductor at stat.math.ethz.ch
>
>
> I am a recent user of the LimmaGUI package. I searched the archives and couldn't
> find this issue raised, so I hope somebody can help.
>
> I'm finding that LimmaGUI works very well to look at what genes are
> differentially expressed, as long as they are expressed to some extent on both
> samples (channels). I am currently mainly interested in extremes of expression:
> looking at genes that are inactive in one sample, but active in the other, and
> I'm finding that LimmaGUI is disregarding all those genes that have no signal
> on one channel (no matter how strong the signal on the other)... so all the
> genes that I want are out of the list of differentially expressed genes.

This is a characteristic of local background correction, it is not specifically a characteristic
of limmaGUI.  It has everything to do with with the choice of image analysis program and
background estimator that you're using.

> How could I get around this?

1. Don't background correct, or
2. Use a background correction option which ensures positive corrected intensities, or
3. Switch to an image analysis program with a low stable background estimator available, e.g, SPOT
or GenePix 6 with morph background.

The authors of limma and limmaGUI always do 2. or 3. in their own work, because of the sort of
problem that you mention.

Gordon

> Jose
>
> --
> Dr. Jose I. de las Heras                      Email: J.delasHeras at ed.ac.uk
> The Wellcome Trust Centre for Cell Biology    Phone: +44 (0)131 6513374
> Institute for Cell & Molecular Biology        Fax:   +44 (0)131 6507360
> Swann Building, Mayfield Road
> University of Edinburgh
> Edinburgh EH9 3JR
> UK