[BioC] topTable vs decideTests

Morten Mattingsdal mortenm at inbox.com
Fri Dec 15 10:58:20 CET 2006

Hi Ingrid,

I think the differences you observe in how your specify the number 
parameter in the topTable function.
Try something like the following code:

top <- topTable(fit2,coef=1, adjust="BH", sort.by="P", 
top <- top[top$adj.P.Val <0.01,]


desTest <- decideTests(fit2, adjust.method="BH", p.value = 0.01)

gives you the same number of DE genes for a given contrast (notice that 
you have to specify contrasts or "coef", in topTable)
If you want to use fold change cutoff of M>2, you can also specify this in:

top <- topTable(fit2,coef=1, adjust="BH", sort.by="M", 
top <- top[top$M>2,]

hope this helps,

Ingrid H. G. Østensen wrote:
> Hi
> I am analyzing Affymetix data and have loaded the data, done gcrma, made the design matrix, lmFit, makeContrasts, contrasts.fit and eBayes. Now I would like to know which genes are differentially expressed and have a fold change >= +-2. 
> I have used the following command to get an overview of the changing genes over all the contrasts:
> desTest <- decideTests(fitContr2, adjust.method=”BH”, p.value = 0.01) 
> This is a great function! But how can I get more information about the genes: p-values etc? I have tried to use topTable 
> tilFil <- topTable(fitContr2, coef = 1, number= 100, adjust.method = “BH”)
> but it is hard to get the same genes from the two functions (topTable and decideTest), and it is hard to find a good cut of for the number of genes to include in the topTable function.
> Does any one have any suggestions?
> When I use topTable I get different parameters and I am trying to understand what the different parameters can tell me but I am still confused about M and log-odds. (And yes I have read a lot of documentation and books, but no one seems to agree/explain it good enough for me to understand) 
> Regards,
> Ingrid :-)
> 	[[alternative HTML version deleted]]
> ------------------------------------------------------------------------
> _______________________________________________
> 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

More information about the Bioconductor mailing list