Question

EdgeR - How to define a group when calculating BCV

0

Entering edit mode

viktoria.c.karlsson • 0

@viktoriackarlsson-22914

Last seen 4.8 years ago

I use edgeR for mRNA analysis and I want to see the difference in BCV between the control group and the treated group. I see that there is an argument called group in the estimateCommonDisp function, but I can not find examples on how to use it. It seems logic to use group=y$samples$group or something like that (where y is my DGElist object) but that does not work. Is there somewhere where I can read about this? I have searched a lot without success.

Best, Viktoria

edgeR • 1.3k views

ADD COMMENT • link updated 4.8 years ago by Yunshun Chen ▴ 900 • written 4.8 years ago by viktoria.c.karlsson • 0

score 1 · Answer 1 · 2020-02-14

1

Entering edit mode

Yunshun Chen ▴ 900

@yunshun-chen-5451

Last seen 8 days ago

Australia

The edgeR pipeline estimates one dispersion (BCV^2) for each gene. It doesn't test for the difference in BCV between groups. The estimateCommonDisp function estimates a "common" dispersion for all the genes. The details of the function and its argument are well documented in its help page, which may be a good place for you to start with.

PS: The estimateCommonDisp function is part of the "classic" edgeR pipeline, which is not generally recommended. We use estimateDisp instead. Please read through the edgeR user's guide for more detail.

ADD COMMENT • link 4.8 years ago Yunshun Chen ▴ 900

0

Entering edit mode

Thanks for your response! I thought that was possible since I saw a comparison between BCV in normal and tumor tissue. I also do not understand this: if I take all samples into consideration when calculating the BCV, wouldn't the value be very affected if there is a big expression difference between the controles and the treated ones? I mean, then the differences wouldn't be due to genetic biological variance, but rather to differences caused by the treatment. Perhaps a stupid thought...

And thanks for your advice regarding the workflow. I will definitely look at it!

ADD REPLY • link 4.8 years ago viktoria.c.karlsson • 0

0

Entering edit mode

I thought that was possible since I saw a comparison between BCV in normal and tumor tissue.

Where did you see testing for differences in BCV between groups using edgeR?

if I take all samples into consideration when calculating the BCV, wouldn't the value be very affected if there is a big expression difference between the controles and the treated ones?

BCV measures the variation between the biological replicates (i.e., samples within the groups). The group information is taken into account when estimating the BCV.

ADD REPLY • link 4.8 years ago Yunshun Chen ▴ 900