I am working on a Proteomics microarray data using only the Red Channel, though there are both R and G channels. The objective is find DE genes in Grade2 samples of cancer as compared to Controls. I created a gist here : Targets file: Volcano Plot: http://share.pho.to/4e1QT I am a bit skeptical about the nature of my volcano plot, showing quite high log odds and skewed. Have I, in the process of playing around with the code, committed a mistake somewhere? Saket

Also, is there a more established way of dealing with Proteomics data using limma? On 28 January 2014 21:09, Saket Choudhary <saketkc at="" gmail.com=""> wrote: > I am working on a Proteomics microarray data using only the Red > Channel, though there are both R and G channels. The objective is find > DE genes in Grade2 samples of cancer as compared to Controls. > > I created a gist here : > Targets file: > Volcano Plot: http://share.pho.to/4e1QT > > > I am a bit skeptical about the nature of my volcano plot, showing > quite high log odds and skewed. Have I, in the process of playing > around with the code, committed a mistake somewhere? > > > > Saket

Any comments on this? Saket On 28 January 2014 21:09, Saket Choudhary <saketkc at="" gmail.com=""> wrote: > I am working on a Proteomics microarray data using only the Red > Channel, though there are both R and G channels. The objective is find > DE genes in Grade2 samples of cancer as compared to Controls. > > I created a gist here : > Targets file: > Volcano Plot: http://share.pho.to/4e1QT > > > I am a bit skeptical about the nature of my volcano plot, showing > quite high log odds and skewed. Have I, in the process of playing > around with the code, committed a mistake somewhere? > > > > Saket

Dear Saket, You are right to be skeptical about the results, because the code is not testing the correct contrast. If you use the group-parametrization as in: design <- model.matrix(~0 + f) then you need to form a contrast between the two groups (Grade 2 vs Control) in order to test for differential expression. As it is, you are simply testing whether the mean for Grade 2 is equal to zero: topTable(ebayes, coef = 2, adjust = "BH", n = 100) and it is no surprise than everything is significant. Section 9.2 of the limma User's Guide explains two different ways you can form the design matrix. Either way is fine, but your code has combined a bit of one approach with a bit of the other. Best wishes Gordon of the > Date: Tue, 28 Jan 2014 21:09:32 +0530 > From: Saket Choudhary <saketkc at="" gmail.com=""> > To: bioconductor at r-project.org > Subject: [BioC] limma and Volcano Plots > Content-Type: text/plain; charset=ISO-8859-1 > > I am working on a Proteomics microarray data using only the Red > Channel, though there are both R and G channels. The objective is find > DE genes in Grade2 samples of cancer as compared to Controls. > > I created a gist here : > Targets file: > Volcano Plot: http://share.pho.to/4e1QT > > > I am a bit skeptical about the nature of my volcano plot, showing > quite high log odds and skewed. Have I, in the process of playing > around with the code, committed a mistake somewhere? > > > > Saket ______________________________________________________________________ The information in this email is confidential and intend...{{dropped:4}}