baySeq

[ksherwood]

Hi,
I would appreciate some help from anyone out there that might be able to assist!

I am using baySeq to analyse the DGE-tag illumina sequenced data. I have also used DESeq and edgeR. I am trying to identify how similar the results using all three methods are.

However I would like to know how to modify the PLOTMA in baySeq? In the simulated data example, the PLOTMA is shown as;

> plotMA.CD(CD, samplesA = 1:5, samplesB = 6:10, col = c(rep("red",
+ 100), rep("black", 900)))

However, the colours are set as red for the 100 data points which have been simulated to be DE ("The data are simulated such that the first hundred counts show differential expression between the first five libraries and the second five libraries.").

Obviously I don't know which data points in my data sets will be DE- how to I produce the plotMA which colours, say, the top 100 most significantly DE genes? I am able to produce the plotMA graph with no colours.

thank you in advance for your help,

[biokari]

I have the same problem, help!!!!!

[dpryan]

The general idea is to sort the p-value/logFC/whatever that you want to color by, select the 100th (or whatever) highest/lowest value from that, and then use ifelse() to do the coloring. The following is some basic R pseudo-code, since I don't use baySeq:

threshold <- sort(abs(logFC),partial=length(logFC)-100, decreasing=T)[length(logFC)-100]
plotMA.CD(CD, samplesA = 1:5, samplesB = 6:10, col = ifelse(abs(logFC) >= threshold, "red", "black"))