[BioC] plotSmear() function in edgeR

Mark Robinson mrobinson at wehi.EDU.AU
Wed Jan 26 10:44:34 CET 2011


Hi Martin.

Thanks for pointing this out.  We should add this detail to the help doc
or manual.

Here is the (somewhat formal) definition. Say X_i is your count for (gene
k from) one sample and Y_k is the count for the other sample, n_X, n_Y is
the library size (or effective library size after calculating
normalization factors).

Now, say X_k=0 and Y_k>0.  logFC = log2( [0/n_X] / [Y_k/n_Y] ) is not
going to work, so we simply replace that with log2( [(0+X*)/n_X] /
[Y_k/n_Y] ) where X* is the minimum count greater than 0.  Often, this
will be 1 and so you are effectively treating 0s as 1s for the
calculation.  Same for both directions.  This is of course just a
visualization and all of these points are shown in the separated smear
part of the plot anyways.

Hope that makes sense.

Cheers,
Mark

> Hi all,
>
>
>
> I have used edgeR to analyse my Illumina RNAseq data. It generally works
> fine and I get biologically meaningfull DGE results.
>
>
>
> I appreciate that plotSmear() is able to visualize the tags with no counts
> in one of the libraries. I am unsure however, as to what value is used to
> plot these tags on the logFC-Axis.
>
>
>
> I have not been able to find this in the edgeR-manual or Mark Robinson's
> /Davis McCarthy's paper. Any ideas on this would be appreciated!
>
>
>
> Best,
>
>
>
> martin
> ______________________________________________________________________________________________________________________
> Martin Umhang
> PhD Student
> D-Biol, ETH Zürich
> LFW, E54
> Universitätstr. 2
> 8092 Zürich
>
> Tel.: 044 6325948 (Lab)
>         044 6323839 (Office)
>
> _______________________________________________
> 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
>



______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}



More information about the Bioconductor mailing list