A WGCNA analysis, need to build a matrix later, take TCGA_AML_Members and consMembers for fisher test

Question

[zzcoliuxuan]

I am doing a WGCNA analysis, and there is a piece of code in it: I need to build a matrix later, take TCGA_AML_Members and consMembers for fisher test:

Numbers of femaleand consensus modules

nTCGA_AML_Mods =length(TCGA_AML_Modules)

nConsMods =length(consModules)

Initialize tablesof p-values and of the corresponding counts

pTable = matrix(0,nrow = nTCGA_AML_Mods, ncol = nConsMods);

CountTbl =matrix(0, nrow = nTCGA_AML_Mods, ncol = nConsMods);

Execute all pairwaisecomparisons

for (Tmod in 1: nTCGA_AML_Mods)

for (Cmod in 1: nConsMods)

{

TCGA_AML_Members =(TCGA_AML_Colors == TCGA_AML_Modules[Tmod]);

consMembers =(moduleColors == consModules[Cmod]);

pTable[Tmod, Cmod]= -log10(fisher.test(TCGA_AML_Members, consMembers, alternative = "greater")$p.value);

CountTbl[Tmod, Cmod]= sum(TCGA_AML_Colors == TCGA_AML_Modules[Tmod] & moduleColors ==consModules[Cmod])

}

Error in fisher. Test(TCGA_AML_Members, consMembers, alternative = "greater") :

how can i do this?

Answer 1

[zzcoliuxuan]

Answer 2

[LChart]

Fisher test needs a 2x2 table. Try fisher.test(table(TCGA, cons), ...). You also may need to use factor(TCGA, levels=(0,1)) for each argument, or else cases where there aren't any overlap will result in a 1x2 table and break the function.