[BioC] Limma: correct calculation of B statistics (log odds)

[J.delasHeras at ed.ac.uk]

Hi Ben,

the only thing that changes is the B value. Everything else (I think!) 
stays unaffected. If you don't want to use the B value, then I think 
you can ignore that parameter (proportion) because I haven't noticed 
any differences in the P values obtained, either adjusted or 
non-adjusted for multiple testing.

Jose


Quoting "Wittner, Ben, Ph.D." <Wittner.Ben at mgh.harvard.edu>:

> Jose,
>
> I'm very glad you asked this question. One of the things that has 
> made me wary
> of using limma is that the proportion of differentially expressed 
> genes is often
> one of the primary things I'm trying to discover from the data, so I 
> feel uneasy
> making an assumption as to what that proportion is. In your email 
> below, you say
> that the output of limma is sensitive to the assumption, which, of 
> course, makes
> me feel even more uneasy about it.
>
> I've not noticed any responses on the BioC list. Has anyone commented on this
> issue to you?
>
> -Ben
>
>> -----Original Message-----
>> From: bioconductor-bounces at stat.math.ethz.ch [mailto:bioconductor-
>> bounces at stat.math.ethz.ch] On Behalf Of J.delasHeras at ed.ac.uk
>> Sent: Wednesday, April 19, 2006 8:06 AM
>> To: bioconductor at stat.math.ethz.ch
>> Subject: [BioC] Limma: correct calculation of B statistics (log odds)
>>
>>
>> I have been using B values to rank genes in order of more likely to
>> less likely (differentially expressed) in LimmaGUI.
>>
>> I am now using Limma, I noticed the default value for the parameter
>> "proportion" (on the function eBayes) is set at 0.01 (expected 1%
>> differentially expressed genes). I didn't pay much attention to this
>> parameter before, because in LimmaGUI you cannot specify it.
>>
>> However, now that I use "straight" Limma more I was playing with the
>> proportion parameter and it affects the B stats a lot. Therefore I come
>> to the question of what's the best way to estimate this parameter.
>>
>> My first guess is to use the P values (FDR, calculated by BH) to decide
>> a cut off, usually 0.05. Then see how many genes are differentially
>> expressed according to that rule. And use this observed proportion of
>> differentially expressed genes as my proportion parameter.
>>
>> Is this the correct way to do it?
>>
>> 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
>>
>> _______________________________________________
>> Bioconductor mailing list
>> Bioconductor at stat.math.ethz.ch
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>> Search the archives:
>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>



-- 
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