[BioC] Limma vennDiagram function

[Lourdes Peña Castillo]

Hello Everyone,

I have encountered an unexpected behaviour in limma vennDiagram function and
I am unsure whether that is the intended behaviour and if so, what the
interpretation is.

Here are the versions I am using:
> sessionInfo()
R version 2.12.0 (2010-10-15)
Platform: i386-apple-darwin9.8.0/i386 (32-bit)

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

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

other attached packages:
[1] limma_3.6.6

loaded via a namespace (and not attached):
[1] tools_2.12.0

I have analysed my data using limma and then I got whether the related
t-statistics are up, down or not significant for three contrasts of
interest:

results <- decideTests(fit2, method = "separate", adjust.method = "none",
p.value = 0.01)

and then I generated a vennDiagram:

## First diagram
vennDiagram(results[,c(1,3,5)],include=c("up","down"),counts.col=c("red","green"))
### this produces the attached image 1 showing no intersection between Comp3
and Comp5.

However, vennCounts gives a different story:

vennCounts(results[,c(1,3,5)])
     Comp1 Comp3 Comp5 Counts
[1,]          0         0           0  17366
[2,]          0         0           1    305
[3,]          0         1           0    141
[4,]          0         1           1    128
[5,]          1         0           0    156
[6,]          1         0           1     40
[7,]          1         1           0     25
[8,]          1         1           1      2

vennDiagram(vennCounts(results[,c(1,3,5)])) ### this produce the attached
image 2

So, image1 shows no intersection between Comp3 and Comp5, but image2 shows
an intersection between Comp3 and Comp5.

My question is shouldn't the numbers agree between both venn diagrams?

Thank you for your clarifications!

Lourdes