Differentially expressed genes across three different conditions
1
0
Entering edit mode
@c727fff3
Last seen 13 months ago
Netherlands

Hello bioconductor team, I would like your guidance about how would you compare gene expression among three different conditions. For example two different tratments and a control group. Hence, levels of tratment are ("control", "treatment_A"," treatment_B") I already tried to run the model with "control" as the reference level:

zlmCond<- MAST::zlm(~treatment + cngeneson + (1|donor_identity), method="glmer", ebayes=FALSE) summaryCond<-summary(zlmCond, doLRT=c("treatment_A", "treatment_B"))

With this I obtain the coefficients and p-values for each tratment (A and B), regarding the control group. But then if I also would like to know the difference between treatment_A" and "treatment_B", Should I run this again but just changing the reference level for one of the treatment groups? Or there is a more "elegant" way to compare simultaneosly the three different groups? What would you suggest?

Thanks a lot for your time and for your cool packages. Best regards, Sofía


# include your problematic code here with any corresponding output
# please also include the results of running the following in an R session

sessionInfo( )

MAST SingleCell diffGeneAnalysis • 298 views
0
Entering edit mode
@james-w-macdonald-5106
Last seen 13 hours ago
United States

See ?ZlmFit-class, particularly the examples for how you fit a contrast between coefficients using lrTest. Also please note that (treatment_A - control) - (treatment_B - control) is the same as treatment_A - treatment_B, so you don't need to do anything extra, just fit an LRT test using the existing coefficients