[BioC] logFC , Limma (microArray Analysis)

[James W. MacDonald]

Hi Cherif,

It would be a good idea to look at the data first. I simply googled your 
GSE number and clicked on one of the samples, and it says this:

*ID_REF* 	
*VALUE* 	Signal intensity value, computed by MAS5. Scaled to 500
*ABS_CALL* 	Detection call
*DETECTION P-VALUE* 	Detection P-value


And if you look at your data you will notice that everything is strictly 
positive, and the maximum is somewhere near 1e14 or so. These are both 
indications that your data have not been logged, which is necessary for 
the analysis of these data.

Best,

Jim


On 4/6/2014 12:48 PM, BEN HAMDA Cherif wrote:
> Hello
> Hope that everything is good for you !
>
> i have a strange results in my work  ( series_matrix.txt microarray
> analysis )
> and i don't know if it's normal or not !
>
> i get high value (logFC)
>
>                logFC    AveExpr         t      P.Value          adj.P.Val
>        B
> 14990  *184.8393  *279.5879  7.053297 6.558774e-08 0.001457032 -3.553707
> 11016 *-113.2735*  132.3939 -6.777861 1.404067e-07 0.001559568 -3.588662
> 583     *565.7128   * 942.2667 -6.602527 2.288115e-07 0.001694349 -3.611925
> 21147 -*103.8393*   411.5121 -6.436832 3.639169e-07 0.002021103 -3.634652
> 2067  -*173.5342*   357.1273 -6.261612 5.958922e-07 0.002647549 -3.659485
> 5404  -*427.6940  * 464.8182 -6.131110 8.616234e-07 0.002968063 -3.678522
>
> i show my method :
>
> gse <- getGEO(filename='GSE1849-GPL96_series_matrix.txt.gz')
>
> i make the design and after that i make analysis with limma :
>
> library(limma)
> fit <- lmFit(exprs(gse), design)
> contrast.matrix <- makeContrasts(Normal_ACS = Normal - ACS , levels=design)
> contrast.matrix
> Ctrl_fits <- contrasts.fit(fit, contrast.matrix)
> Ctrl_ebFit <- eBayes(Ctrl_fits)
> nrow(topTable(Ctrl_ebFit, adjust = "BH", coef = 1, number = Inf)
>
> So it's normal or not ? i made mistake in my R code ? someone can help me ?
>
> Thank you so much
>
> Best,
> cherif
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, M.S.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099