How to properly average normalized counts of biological replicates
1
3
Entering edit mode
Marina ▴ 40
@marina-15019
Last seen 3.8 years ago

Dear all, I'm trying to combine the normalized counts from 3 biological replicates into single columns of the normalized count matrix. I need it in order to "absolutely" analyze the expression of specific genes, for example - presenting expressions on heat maps (or other plots) and doing co expression analysis. I would like to emphasize that all 3 biological replicates are, in my point of view, identical, meaning, all are before any treatment, or all of them are after a specific treatment. Yet, I cannot treat them as technical replicates as they were taken from different biological samples. I know that Deseq2 can collapse technical replicates into a single column, but it is not intended for biological replicates (and I must say that I'm not fully understand why, if all "averaged" replicates represent the same treatment but I'm still a newbie and I accept the verdict). What alternatives do I have? Is it allowed to do a simple average on the normalized counts of each of the 3 samples converting it into a single normalized count?? Any suggestion would be highly appreciated.

DESeq2 Normalization RNASeqR • 5.1k views
ADD COMMENT
1
Entering edit mode
@mikelove
Last seen 20 hours ago
United States

If you want to calculate the average normalized count for a group of samples:

ave.cnt.A <- rowMeans( counts(dds, normalized=TRUE)[ , dds$condition == "A" ] )
ADD COMMENT
1
Entering edit mode

Thank you so much!! I'll try that. If I understand correctly then a simple average of the 3 (for example) normalized count columns is what I need to do? For some reason, I had in my mind that a more careful averaging was required, something that takes into consideration some additional statistical properties, but I have no idea why I was thinking this way...

ADD REPLY
0
Entering edit mode

"is what i need to do"?

Yes to calculate the average scaled count, the above is the code you would use.

ADD REPLY
0
Entering edit mode

Hi if I have 12 samples with 3 biological replicates ( 36 sample total), and I want to create a matrix with row names as gene and columns as the average ( average of 3 replicate) normalized counts of 12 sample, how may I create that in R? TIA.

ADD REPLY
0
Entering edit mode

Hi if I have 12 samples with 3 biological replicates ( 36 sample total), and I want to create a matrix with row names as gene and columns as the average ( average of 3 replicate) normalized counts of 12 sample, how may I create that in R? TIA.

ADD REPLY
0
Entering edit mode

You can do this with the above code, plus a loop over the samples. I'd recommend getting some help from someone familiar with basic R scripting.

ADD REPLY
0
Entering edit mode

Thank you.

ADD REPLY
0
Entering edit mode

Curious if taking an average (mean) of (3) biological replicates if a best way to represent each sample data? is there any other statistical analysis that needs to be taken into account before taking average (mean), median or other?

ADD REPLY
1
Entering edit mode

This is just for descriptive summary, so mean of scaled counts is fine.

For statistical testing, we model the counts and their dispersion.

For plotting on a heatmap or otherwise, we recommend variance stabilization with vst().

ADD REPLY
0
Entering edit mode

Hi please let me know how to implement this code, specially [ dds$condition == "A"] this part. how it should be changed ?

ADD REPLY
0
Entering edit mode

What is =="A" here?

ADD REPLY

Login before adding your answer.

Traffic: 789 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6