Featurecounts normalize by gene length

Question

[David]

Hello,

I have run featureCounts on aligned bam files (DNA from bacterial genomes).

I have a table with counts per sample as expected. However i would like to normalize by gene length since genes of different length can have more reads mapped then genes that are shorter.

I was wondering what would be the best way to do that ?? Any tool or script ? Once i have normalized by gene length i would like to feed it to DESEq2

Thanks

Answer 1

[ATpoint]

DESeq2 takes care of all necessary steps internally. There is no need to manipulate the data prior to creating your DESeq2 object. Raw counts are expected. Please read the manual.

[David]

Hi , I understand deseq will normalize by library size but prior to that you need to normalize by gene length. Am i correct?

[ATpoint]

Not really. There is no need to account for gene length because gene length is identical for all conditions. See also as a starting point for further online searches: https://support.bioconductor.org/p/67131/ and DESeq2 normalisation: is the size of the gene taken into account?

[David]

Ok thanks, So ranking genes by expression in this context should not be possible.

[WouterDeCoster]

How do you mean it's not possible? DESeq will do a far more sophisticated normalization than you can think off, and you can get the normalized counts. Then you can rank whatever you want. And now, before we help you any further, you are going to read all the links ATpoint gave to you.