How to calculate Differential Gene expression between samples with replicates

Question

[doncheng1]

I know this might sound silly to some of you, I am quiet new to this.

I have some RPKM data for groups of samples, one is a control and the other is the wild type. Each group has 3 replicates. I would like to calculate the differential expression for the genes between the two samples. How do I then treat the replicates within each group when calculating the differential expression.

Thanks in advance

Answer 1

[Benn]

Did you check any previous posts? There are many similar posts!!

Answer 2

[Biogeek]

EdgeR / Deseq/ Limma-Voom are your best bets for R software.

Raw counts input, not RPKM! as these programs do the normalisation themselves based on library size.

Negative binomial GLM in EdgeR is quite a popular method, I'd recommend you read the EdgeR vignette and learn how to use R. It's rewarding once you know how. If you are starting out, maybe the Trinity pipeline is your best bet. It's got heaps support.

Answer 3

[Charles Warden]

1) If you use log2(RPKM + X), where X is a rounding threshold like 0.1 or 1, then you can use standard methods for roughly normally distributed data

2) Depending upon how RPKM was calculated, there is probably a column/table of estimated counts, which you can use in edgeR/DESeq/limma-voom, etc. If the abundances are for transcripts, then you should probably add the estimated counts for transcripts for the same gene, for gene-level differential expression.