[BioC] TLDA data

Heidi Dvinge heidi at ebi.ac.uk
Wed Jun 16 19:39:29 CEST 2010

Hello Deepak,

> Dear sir
> Almost all analysis are working well. Thank you for all your helps
> throughout the process. HTqPCR package was very help full because i
> was able generate heat maps and performed clustering analysis. Now i
> am having few dought which i didt find in the vintage also. What is
> iqr.value and for what this is done.
The IQR value is the interquartile range. It measures how much  
variation you see across your data for each gene, so it's a way for  
you to remove genes that show only little or almost no variation  
across your samples, since these are unlikely to be differentially  
expressed anyway. For high-content microarrays removing these genes  
is an advantage, since removing "uninteresting" genes before the  
downstream analysis means that when you correct your p-values for  
multiple testing, this correction will be less strict.

For qPCR cards the number of genes is typically quite low, so this is  
less of an issue. Removing the low-IQR genes prior to the analysis of  
differential expression probably doesn't make much difference for  
your p-values.
>   For me when i have done t test
> it showed some error but mann whitney test worked. After mann whitney
> test i got fold change mean value of all sample in a tumor group
> compared to normal group. Can we get fold change value of individual
> samples in tumor group compared to mean of normal samples.
I'm afraid that this is something you'd have to calculate "manually",  
i.e. using non-HTqPCR functions. It shouldn't be too difficult  
though. If you e.g. have 20 samples and a qPCRset names "q", and the  
first 10 are test and the the last 10 are controls, you can say e.g.

data <- getCt(q)
data.test <- data[,1:10]
data.control <- data[,11:20]
mean.control <- rowMeans(data.control)
fc.test <- data.test - mean.control

> Could we
> get result in excel or as separate sheet as we get plots.
In R you can always write results to a file using for example  
write.table. To get a tab-separated excel-like file, say:

# Load example preprocessed data
# Test between two groups, collapsing replicated features
diff.exp <- mannwhitneyCtData(qPCRpros[,1:4], groups=factor(c("A",  
"B", "B", "A")), calibrator="B")
# Write to file
write.table(diff.exp, file="test.xls", sep="\t", quote=FALSE,  
> Once again thank you for your help during the analysis.
You're welcome.
> > iqr.values <- apply(exprs(d.norm), 1, IQR)
> Error in quantile.default(as.numeric(x), c(0.25, 0.75), na.rm =  
> na.rm,  :
>  missing values and NaN's not allowed if 'na.rm' is FALSE
As the error says, this doesn't work if you have NAs in your data. If  
you look at the help file for IQR, ?IQR, you'll see that there's an  
option na.rm which you can set to TRUE. When using the apply  
function, this would be:
qr.values <- apply(exprs(d.norm), 1, IQR, na.rm=TRUE)

> > qDE.ttest <- ttestCtData(sr.norm[,1:4], groups=files$Treatment[1:4],
> + calibrator="Normal")
> Error in ttestCtData(sr.norm[, 1:4], groups = files$Treatment[1:4],
> calibrator = "Normal") :
>  Two factor levels required for 'groups
Do the first 4 samples actually belong to different groups?? I.e. are  
there two different groups listed in files$Treatment[1:4]? Otherwise  
you're trying to perform a t-test of all samples against nothing,  
which isn't possible.

> > sessionInfo()
> R version 2.11.0 (2010-04-22)
> x86_64-pc-mingw32
> locale:
> [1] LC_COLLATE=English_India.1252  LC_CTYPE=English_India.1252
> LC_MONETARY=English_India.1252 LC_NUMERIC=C
> LC_TIME=English_India.1252
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
> other attached packages:
> [1] HTqPCR_1.2.0       limma_3.4.1        RColorBrewer_1.0-2  
> Biobase_2.8.0
> loaded via a namespace (and not attached):
> [1] affy_1.26.1           affyio_1.16.0         gdata_2.8.0
> gplots_2.7.4          gtools_2.6.2          preprocessCore_1.10.0
> tools_2.11.0
> >
> -- 
> Deepak Roshan V G
> Laboratory Of Cell Cycle Regulation  &   Molecular Oncology
> Division of Cancer Research
> Regional Cancer Centre
> Thiruvananthapuram
> Kerala, India 695 011

More information about the Bioconductor mailing list