Column-wise operations based on column-name prefixses

Question

Please consider:

require(dplyr)

my_data <- 
  data.frame(a.x = sample(1100),
             b.x = sample(1100),
             a.y = sample(1100),