[BioC] LIMMA makeContrast one test and two controls

Robert Ivanek ivanek at biomed.cas.cz
Fri Nov 18 18:27:36 CET 2005


Hello,

I have a question regarding the contrasts.matrix.
This is design of my experiment:

SlideNumber	FileName	Cy3	Cy5
23	Mo_23.gpr	pool	con1
24	Mo_24.gpr	pool	con2
25	Mo_25.gpr	pool	test
26	Mo_26.gpr	pool	con1
27	Mo_27.gpr	pool	con2
28	Mo_28.gpr	pool	test
15	Mo_15.gpr	pool	con2
16	Mo_16.gpr	pool	con2
17	Mo_17.gpr	pool	con1
18	Mo_18.gpr	pool	con1
19	Mo_19.gpr	pool	con2
20	Mo_20.gpr	pool	con2
21	Mo_21.gpr	pool	test
22	Mo_22.gpr	pool	test
29	Mo_29.gpr	pool	con1
30	Mo_30.gpr	pool	con1
31	Mo_31.gpr	pool	con2
32	Mo_32.gpr	pool	con2
33	Mo_33.gpr	pool	con1
34	Mo_34.gpr	pool	con1
35	Mo_35.gpr	pool	con2
36	Mo_36.gpr	pool	con2
37	Mo_37.gpr	pool	test
38	Mo_38.gpr	pool	test
02	Mo_02.gpr	pool	test
03	Mo_03.gpr	pool	test
05	Mo_05.gpr	pool	con1
06	Mo_06.gpr	pool	con1
39	Mo_39.gpr	pool	test
40	Mo_40.gpr	pool	test


It is common reference design, with reference in Cy3 channel, called
"pool".

I would like to compare the "test" vs. both controls "con1" and
"con2". What would be the syntax of makeContrast?

design <- modelMatrix(targets, ref="pool")
contrast.matrix <- makeContrasts(Test.vs.Controls=(test-(con1+con2)/2),levels=design)   
contrast.matrix
     Test.vs.Controls
     con1             -0.5
     con2             -0.5
     test              1.0
     
Is this correct?

Thanks for your help


-- 
Robert Ivanek



More information about the Bioconductor mailing list