Skip to main content
Topic: compare more than 2 groups (Read 3224 times) previous topic - next topic

compare more than 2 groups

Hi,

I was wondering whether it is possible to compare more than 2 groups. I have one "test" group and two control groups that I would like to compare. I am most interested in features that differ between the "test group" and both control groups. Now I manually check many EICs to see whether this is the case. It would be somewhat easier if I could have diffreport select these features for me. Extracting these features from the results file generated by "diffreport" is of course relatively easy, but I would then still get many box plots and EICs of features in which I am not so much interested.

Koen

 

Re: compare more than 2 groups

Reply #1
Hi,

not exactly sure what you mean there.

Do you want the features differential between the test group
and both control groups ? Then you could assign a single sampclass
to both controls. Or are you interested in the intersection of features
differential between test and both (or either of ?) the controls ?

You can do any logical set operations if you create several
individual diffreports between class1 and class2 of your choice(s),
and then combine them with and/or: dr1$pvalue>0.05 & dr2$pvalue>0.05

Make sure the individual diffreports are unsorted using sortpval=FALSE,
so that a feature in row 17 is the same across the diffreports.

Yours,
Steffen
--
IPB Halle                          Mass spectrometry & Bioinformatics
Dr. Steffen Neumann         http://www.IPB-Halle.DE
Weinberg 3 06120 Halle     Tel. +49 (0) 345 5582 - 1470
sneumann(at)IPB-Halle.DE