DESeq2 FPKM normalization
2
1
Entering edit mode
ribioinfo ▴ 90
@ribioinfo-9434
Last seen 13 months ago

Hi if I use the FPKM I can compare the expression across different samples and different experiments.

With DESeq2 I can compare the expression of the genes that are in the normalized table. How can I use the DESeq2 normalization and to compare the expression of a gene in samples that are in different analyses?

Thanks

deseq2 normalization • 11k views
ADD COMMENT
1
Entering edit mode
@wolfgang-huber-3550
Last seen 7 weeks ago
EMBL European Molecular Biology Laborat…

The DESeq normalisation is intended for relatively precise quantitative comparisons of samples that were consistently processed in the same experiment or study. It is not intended for comparisons across heterogeneous experiments/studies.

Methods for the latter include FPKM, TPM.

But note that such comparisons then tend to be of a more qualitative nature. Because of 'batch effects', it could require a lot of statistical finesse to meaningfully apply e.g. ahypothesis test or a  (generalised) linear model approach. Nothing is impossible, but certainly it's not easy.


 

ADD COMMENT
0
Entering edit mode

So if we need to compare between two sets of data from two different experiments, is it worth using COMBAT to remove batch effect and then use edgeR to normalize by TPM and DE analysis?

ADD REPLY
0
Entering edit mode
ribioinfo ▴ 90
@ribioinfo-9434
Last seen 13 months ago

Thank you. I have found this post: /Using DESeq normalized gene count to replace FPKM?

where there is this sentence:

"If you want to have the same scale as typical FPKM values (and so have better comparability across experiments), you could then divide everything by something like geometric mean of the total read counts of all samples / 1 million"

But why have I to use the geometric mean of the total read counts of all samples? If I want something like the FPKM should I not divide each sample by the sum of its read counts?

Thank you

ADD COMMENT
0
Entering edit mode
The column sum is not a robust estimator for the sequencing depth. But you can choose this option in fpkm() if you prefer.
ADD REPLY

Login before adding your answer.

Traffic: 215 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