Question: Using Mann-Whitney U Test to compare replicates
gravatar for jjrin
14 days ago by
jjrin10 wrote:

Hello, I have several replicates of a certain variable (at least 5). I would like to compare these distributions to see how similar they are with each other. I have TPMs as a measurement as well as counts (if necessary). They are not normally distributed and I read that the Mann-Whitney U-Test is an appropriate statistical test. I read about this in the illumina replicate manual but it there are no real explanation as to how this applies or can be used. I tried running it with R (using wilcox.test) but I continuously get p-values above 0.05.

I am also unsure as to how to compare multiple replicates, I am currently only comparing two of the 5 replicates. Is there another method I could use to compare these distributions to see their similarities and differences (preferably with a statistical test that gives me some sort of number and p-value)? Or is there a way to improve how I'm running my U-Test?

Thank you!

This is the code I am using to compare my two out of five replicates:

 wilcox.test(TPM~REPLICATE, data=test)

In my data sheet I have one column with all of the TPMs of the two replicates and in the REPLICATE column I have "1" indicating the TPM is from the first replicates and "2" for the second replicate.

statistics rna-seq R • 150 views
ADD COMMENTlink modified 14 days ago by Jean-Karim Heriche13k • written 14 days ago by jjrin10
gravatar for Jean-Karim Heriche
14 days ago by
EMBL Heidelberg, Germany
Jean-Karim Heriche13k wrote:

What I think you're looking for is an omnibus test, in your case the Kruskal-Wallis test. If you end up rejecting the null hypothesis then you can follow up with a post-hoc analysis using the Mann-Whitney U test and a correction for multiple testing.

ADD COMMENTlink written 14 days ago by Jean-Karim Heriche13k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 599 users visited in the last hour