[BioC] calcNormFactors gives me NaN

Ryan rct at thompsonclan.org
Thu Dec 19 01:13:33 CET 2013


NaN means "not a number", which is used as the result of certain 
arithmetic operations whose results are not well defined, such as 0/0. 
Perhaps when too many genes have all zero counts, edgeR's TMM method 
ends up including some of these, and their fold changes are computed as 
0/0, which yields NaN? What is the result of "summary(d$counts)"? Is 
there someing odd about these counts, such as many genes (rows) having 
all zeroes?

On Wed Dec 18 15:02:46 2013, Audra [guest] wrote:
>
> So, here is my problem.  I have been working with edgeR and normalizing data.  I have successfully done this using calcNormFactors() before with some other RNAseq count data.  Currently I have a set of count data.  Each replicate is saved into a separate tab delimited file with the first column being the gene name and the second being the counts.  I created a target file which contains the file names, groups (time points), and descriptions in that order. I first read in the data using:
>> targets <- read.delim("/users/Audra/desktop/targets.txt")
>> d <- readDGE(targets)
>
> I then normalizing using
>
>> d <- calcNormFactors(d)
>
> However, when I do this, all of my norm.factors come up as NaN.  I am following the same method as I did with the other data set that worked.  Does anyone have any suggestions? I guess I just am not understanding what the NaN means totally.  Is it an error, does it have to do with my library sizes, or have I done something wrong. I have pasted a copy of the work session below.  Thanks for the help!
>
>> library(edgeR)
> Loading required package: limma
>> targets <- read.delim("/users/Audra/desktop/targets.txt")
>> d <- readDGE(targets)
> dim(d)
>> dim(d)
> [1] 50770    38
>> calcNormFactors(d)
> <SNIP>
>
>   -- output of sessionInfo():
>
>> sessionInfo()
> R version 3.0.2 (2013-09-25)
> Platform: x86_64-apple-darwin10.8.0 (64-bit)
>
> locale:
> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] edgeR_3.4.2  limma_3.18.6
>
> loaded via a namespace (and not attached):
> [1] tools_3.0.2
>
> --
> Sent via the guest posting facility at bioconductor.org.
>
> _______________________________________________
> 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



More information about the Bioconductor mailing list