rankToContigTab {wrMisc}

Contingenty tables for fit of ranking

Description

Count the number of instances where the corresponding columns of 'dat' have a value matching the group number as specified by 'grp'. Counting will be performed/repeated independently for each line of 'dat'. Returns array (1st dim is rows of dat, 2nd is unique(grp), 3rd dim is ok/bad), these results may be tested using eg fisher.test. This function was made for prearing to test the ranking of multiple features (lines in 'mat') including replicates (levels of 'grp').

Usage

rankToContigTab(dat, grp)

Arguments

dat

(matrix or data.frame of integer values) ranking of multiple features (lines), equal ranks may occur

grp

(integer) expected ranking

Value

array (1st dim is rows of dat, 2nd is unique(grp), 3rd dim is ok/bad)

See Also

lm

Examples

# Let's create a matrix with ranks (equal ranks do occur)
ma0 <- matrix(rep(1:3,each=6), ncol=6, dimnames=list(
  c("li1","li2","ref"), letters[1:6]))
ma0[1,6] <- 1                       # create item not matching correctly
ma0[2,] <- c(3:1,2,1,3)             # create items not matching correctly
gr0 <- gl(3,2)                      # the expected ranking (as duplicates)
(count0 <- rankToContigTab(ma0,gr0))
cTab <- t(apply(count0, c(1,3) ,sum))
# Now we can compare the ranking of line1 to ref ...
fisher.test(cTab[,c(3,1)])          # test li1 against ref
fisher.test(cTab[,c(3,2)])          # test li2 against ref