I have a data set where samples are grouped according to their health status (3 levels) and sampling timepoints (2 levels for each sample). This would be a simplified version of the data set:
df <- data.frame(status = c( rep("healthy", 4), rep("condition1", 4), rep("condition2", 4) ), time = rep(c("1", "2"), 6), animal = paste0("animal", unlist(lapply(1:6, function(x) rep(x,2)))))
I am interested in detecting differentially expressed genes at each time point between sick and healthy individuals. I am not interested in finding changes over time.
According to the vignette is valid to model differential expression by grouping variables:
df$groups <- paste0(df$status, "time", df$time) model.matrix(~0+groups, df)
I want to do the following 4 contrasts:
groupscondition1time1 vs groupshealthytime1 groupscondition2time1 vs groupshealthytime1 groupscondition1time2 vs groupshealthytime2 groupscondition2time2 vs groupshealthytime2
My question is: should I correct for multple testing across contrasts?
The vignette indicates the following:
"Regarding multiple test correction, if a user is planning to contrast all pairs of many levels, and then selectively reporting the results of only a subset of those pairs, one needs to perform multiple testing across contrasts as well as genes to control for this additional form of multiple testing"
This is not exactly my case, as I have planed the contrasts, but I want to make sure.
Thanks in advance!