Search
Question: Representative gene expression value in one condition with several replicates
0
gravatar for Jack
6 months ago by
Jack0
Jack0 wrote:

Hi all,

I want to know how to get a gene expression value for a condition with different replicates.

For example, I have condition M and N, each condition with two replicates M1, M2, N1, N2

I want to get one value to represent the gene expression value (FPKM or TPM) of M, can I just use the mean of each replicate? M=(M1+M2)/2?

Is there any other way to calculate the gene expression value for a condition?

Any advice is appreciated!

 

 

ADD COMMENTlink modified 6 months ago by Gordon Smyth33k • written 6 months ago by Jack0
5
gravatar for Aaron Lun
6 months ago by
Aaron Lun19k
Cambridge, United Kingdom
Aaron Lun19k wrote:

As Mike says, this isn't an edgeR question. But I will pretend it is. If you have the counts, go through an edgeR analysis - or at least to calling glmFit - with the following design matrix:

group <- c("M", "M", "N", "N")
design <- model.matrix(~0 + group)

You didn't specify the nature of your replicates, but you may need to add a blocking factor if M1 is related to N1 (e.g., from the same individual) and M2 is related to N2.

Anyway, once you've done that, you can obtain the log-average expression of each level of group from the $coefficients field of the output of glmFit. This provides a general approach to getting condition-specific expression values, taking advantage of NB GLMs to give a more precise estimate than averaging FPKMs.

ADD COMMENTlink modified 6 months ago • written 6 months ago by Aaron Lun19k

Thank you very much!!

ADD REPLYlink written 6 months ago by Jack0
4
gravatar for Gordon Smyth
6 months ago by
Gordon Smyth33k
Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia
Gordon Smyth33k wrote:

If you want expression values on a log-scale, then you can use the process explained by Aaron, which is similar to but better than just averaging the individual log-expression values.

If you want expression values on the unlogged scale, then the edgeR package provides functions to do this. Type

library(edgeR)
?cpmByGroup
?rpkmByGroup
ADD COMMENTlink modified 6 months ago • written 6 months ago by Gordon Smyth33k

Thank you very much for you advice!

ADD REPLYlink written 6 months ago by Jack0
0
gravatar for Michael Love
6 months ago by
Michael Love18k
United States
Michael Love18k wrote:

This isn't a DESeq2 (or edgeR) question really, so I'm removing the DESeq2 tag. The arithmetic or geometric mean of the TPM seems to be a reasonable number for the average relative abundance. I don't have any strong opinions about this though.

ADD COMMENTlink modified 6 months ago • written 6 months ago by Michael Love18k

Yes, you are right. I think it is good to hear your opinion.

ADD REPLYlink written 6 months ago by Jack0
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 321 users visited in the last hour