edgeR: number of DGE genes and FDR??

[SpreeFu]

Hi all,

In my concept, if the number obtained from "summary(de<-decideTestsDGE(df, adjust.method="BH", p.value=0.05))" is 7, then there should be 7 genes with FDR less than 0.05 in the list, right?
When I tried to see DGE by edgeR, it present as following:

> summary(de<-decideTestsDGE(df, adjust.method="BH", p.value=0.05))
[,1]
-1 2
0 340
1 5

Here means that the number of differentially expressed gene is 7 (FDR<0.05), right?

But when I wanted to see the top 20 genes, their FDR are all less than 0.05.

> topTags(A_B,20)
Comparison of groups: A-B
logFC logCPM PValue FDR
ATxxx 2.4277458852 11.707828 2.177741e-11 7.556760e-09
ATxxx 2.1995419428 7.696531 4.090971e-06 7.097835e-04
ATxxx 1.3143387328 12.547087 3.155543e-05 3.364114e-03
ATxxx -1.4263203424 11.251632 3.877941e-05 3.364114e-03
ATxxx 1.4985278106 11.174781 9.803395e-05 6.803556e-03
BRODxx -3.0514440353 6.191085 1.635781e-04 9.460266e-03
ATLANxx 0.8367554361 13.197904 4.059972e-04 2.012586e-02
HELxx 1.5246704658 7.394363 2.432827e-03 1.055239e-01
ATCxxx 0.8184084803 12.005165 3.088130e-03 1.190646e-01
ATRxxx 2.3828224280 5.805848 3.581865e-03 1.242907e-01
Sadxxx -1.4732584469 7.605255 5.243566e-03 1.605100e-01
ATxxx -5.0018389997 4.633235 5.550776e-03 1.605100e-01
.............................

Did I misunderstand something?

[bruce01]

You misunderstood scientific notation:

1.055239e-01 = 0.1055239

So there are 7 genes with FDR < 0.05 (fifth column).

[SpreeFu]

You misunderstood scientific notation:

1.055239e-01 = 0.1055239

So there are 7 genes with FDR < 0.05 (fifth column).

Thanks for answering my stupid misunderstood!

[bruce01]

No problem, remember it like this:

1.04e-01 means move the 1 one space back behind the decimal point = 0.104
1.04e-02 means move the 1 two spaces =0.0104

And so on. I still do that in my head every time I see sci. notation!

[SpreeFu]

No problem, remember it like this:

1.04e-01 means move the 1 one space back behind the decimal point = 0.104
1.04e-02 means move the 1 two spaces =0.0104

And so on. I still do that in my head every time I see sci. notation!

Thanks again!

I had other 3 comparisons, but none of them list their FDR with sci. notation. Is that means they are too big and not necessary to use sci. notation, e.g. 0.000745?