limma camera function Multiple contrasts
1
0
Entering edit mode
gaby8638 ▴ 20
@gaby8638-11595
Last seen 4.1 years ago
Mexico, LANGEBIO CINVESTAV, Irapuato

Hi all!!

I am making a differential expression analysis of microarray data with two contrasts. My design matrix is like this

  A_1  A_2    B
   1      0        0
   1      0        0
   1      0        0
   0      1        0
   0      1        0
   0      1        0
   0      1        0
   0      0        1
   0      0        1
   0      0        1
   0      0        1
   0      0        1
   0      0        1
   0      0        1

 

And my contrast matrix, made with the makeContrast function is this

       Contrasts
Levels  A_1-B   A_2 - B
   A_1     1       0
   A_2     0       1
   B      -1      -1

 

I want to find the GO_terms associated to both contrasts, for that I am using the camera function like this

camera(Exprs, Index, design, contrast = contMatrix) 

Where:

Exprs is my matrix of log-expression values 

Index is the list of GO terms associated to the genes in the Exprs matrix

design is the design matrix I showed you before 

and as contrast I input my contMatrix object i showed you before.

 

My doubt is if camera output is actually the GO_terms taking account the TWO contrasts defined. 

If not, how can I get that result.

Thanks!!

limma microarray contrast camera • 1.3k views
ADD COMMENT
0
Entering edit mode
@gordon-smyth
Last seen 8 hours ago
WEHI, Melbourne, Australia

camera() only uses one contrast. I wonder what you expected it to do with two contrasts? Did you expect a test by F-statistic? Or did you expect two separate analyses?

When in doubt it's not a bad idea to read the documentation. The help page for camera() says that the contrast argument is:

contrast of the linear model coefficients for which the test is required. Can be an integer specifying a column of design, or else a numeric vector of same length as the number of columns of design.

As you can see, the documentation says that the contrast can be either an integer or a vector. It cannot be a matrix.

ADD COMMENT
0
Entering edit mode

Thank you Gordon,

Yes I got confused with the documentation...

What I want is the GO terms of both contrast, I think yes like a F-statistic. Do you know any library or function for that?

Thanks before hand,

Gaby

ADD REPLY
0
Entering edit mode

You could do an F-statistic topTable in limma, then pass the DE results to any gene ontology analysis package, such as gostats or goseq.

Alternatively, sticking to the limma package, you could run geneSetTest with the F-statistics for each GO set separately.

I haven't provided this functionality in camera() because camera() is designed to work with directional pathways and F-statistics are not directional. camera() is not designed to work with GO either, because of its redundancy and lack of directionality.

ADD REPLY

Login before adding your answer.

Traffic: 202 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6