Correlation/p values of all combinations of all rows of two matrices

Question

I would like to calculate the correlation and the p value of that correlatio of each species (bac) to each of the factors (fac) in a second data frame. Both were measured at

Answer 1

You can just pass the full matrices to the cor function (or psych::corr.test)and it takes care of finding the correlation of the relevant columns.

For example

cor(t(fac), t(bac))
#            bac1        bac2        bac3        bac4        bac5
# fac1  0.9899495 -0.07559289 -0.60000000 -0.60000000 -0.07559289
# fac2  0.9899495 -0.07559289 -0.60000000 -0.60000000 -0.07559289
# fac3 -0.3207135  0.94285714 -0.07559289 -0.07559289 -0.48571429
# fac4 -0.8000000 -0.32071349  0.98994949  0.98994949 -0.32071349
# fac5 -0.3207135 -0.48571429 -0.07559289 -0.07559289  0.94285714
# fac6         NA          NA          NA          NA          NA

You can then turn this in to long format using reshape2::melt

reshape2::melt(cor(t(fac), t(bac)))
#    Var1 Var2       value
# 1  fac1 bac1  0.98994949
# 2  fac2 bac1  0.98994949
# 3  fac3 bac1 -0.32071349
# 4  fac4 bac1 -0.80000000
# ---
# ---

To get the p-values use the same approach

test <- psych::corr.test(t(fac), t(bac), adjust="none")

And melt as before and join

merge(melt(test$r, value.name="cor"), melt(test$p, value.name="p-value"), by=c("Var1", "Var2"))
#   Var1 Var2         cor    p-value
# 1 fac1 bac1  0.98994949 0.01005051
# 2 fac1 bac2 -0.07559289 0.92440711
# 3 fac1 bac3 -0.60000000 0.40000000
# 4 fac1 bac4 -0.60000000 0.40000000
# 5 fac1 bac5 -0.07559289 0.92440711
# 6 fac2 bac1  0.98994949 0.01005051