Hi all,

I am using LIMMA package to analyse RNA-Seq data from a time course experiment. We have 3 time point, one control, two treatments and all the samples are in triplicate. I used the spline method as suggested in the tutorial so I have:

yS<-DGEList(counts=limma.reads.counts)
SplineTargets=data.frame(File=sampleFileNames,
 ​Group=c(rep("Control",9),rep("Ofx",9),rep("MitC",9)),
 Time=c(rep(0,3),rep(30,3),rep(180,3),rep(0,3),rep(30,3),
      rep(180,3),rep(0,3),rep(30,3),rep(180,3)))
X <- ns(SplineTargets$Time, df=2)
Group <- factor(SplineTargets$Group)
spline.design <- model.matrix(~Group*X)
v <- voom(yS,spline.design,plot=FALSE)
fitSpline <- lmFit(v,spline.design)
fitSpline <- eBayes(fitSpline)

To extract the values of interest:

data.as.top.table = topTable(fitSpline,coef=6:9,n=500)

In the table result I have the columns

object..a..     GroupMitC.X1    GroupOfx.X1     GroupMitC.X2
GroupOfx.X2     AveExpr F       P.Value adj.P.Val

Could you please help me in understanding the meaning of GroupMitC.X1,
GroupOfx.X1, GroupMitC.X2, GroupOfx.X2 columns? How can I interpret
this values? How are they associated with the behaviour of the gene?

Thanks in advance,

Stefania

 -- output of sessionInfo():

R version 3.0.1 (2013-05-16)
Platform: x86_64-apple-darwin10.8.0 (64-bit)

locale:
[1] it_IT.UTF-8/it_IT.UTF-8/it_IT.UTF-8/C/it_IT.UTF-8/it_IT.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] tools_3.0.1

Dear Stefania,

You have a replicated time course experiment with three time points. This is dealt with in detail in the limma User's Guide in Section 9.6.1 "Replicate time points".  You should be able to apply the advice in that section to your data without any problems.

The limma User's Guide tells you that spline curves are only for time course experiments with many time points.  You don't have many time points -- only three.  There is no advantage in using spline curves for a time course experiment with only three time points.  You cannot achieve anything that the simpler analysis presented in Section 9.6.1 would not achieve more explicitly.

Given that you have fitted spline curves, the test that you have conducted will correctly select genes whose change over time is different between the three genotypes.  However you cannot interpret the individual logFC columns in the toptable.  These are simply mathematically defined basis vectors for the spline curve -- they have no meaningful interpretation as individual columns.

Best wishes
Gordon

Note: This answer was originally posted 18 April 2014: Limma table results (time course analysis)