Bulk RNA seq analysis on multiple conditions (4)

Question

[SO-YEON]

Hi guys,

I have bulk RNA sequencing data from samples treated with 4 different concentrations of a drug. I am trying to figure out whether there is a concentration dependent effect on gene expression. However, I'm not sure if this is possible because I find that a lot of RNA seq analysis involves pair-wise comparisons. I guess I can do pair-wise comparisons but I'm not sure how I would put all that data together, if this would show the concentration-dependent effect, and if there will be problems with statistics. I would really like to use ANOVA but it doesn't really seem like it's possible. I was wondering if a package out there that looks at this or if anyone has any experience with such data set? Thank you so much in advance!

[GenoMax]

Someone will provide a more rounded answer but until then check:

https://www.bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.html#time-series-experiments
https://hbctraining.github.io/DGE_workshop_salmon_online/lessons/08b_time_course_analyses.html

Answer 1

[Gordon Smyth]

limma, edgeR and DESeq2 all do ANOVA quite easily. See the User Guides. In the case of limma, you would use:

design <- model.matrix(~Concentration)
fit <- voomLmFit(y, design)
fit <- eBayes(fit)
topTable(fit, coef=2:4)

which outputs ANOVA F-tests.