R to recode variables if the categorical variable's frequency lower than an defined value

Question

Here is an example for the dataset (d):

rs3 rs4 rs5 rs6
1   0   0   0
1   0   1   0
0   0   0   0
2   0   1   0
0   0   0   0
0   2   0   1
0   2   NA  1
0   2   2   1
NA  1   2   1

To check the frequency of the SNP genotype (0,1,2), we can use the table command

table (d$rs3)

The output would be

0 1 2 
5 2 1

Here we want to recode the variables if the genotype 2's frequency is <3, the recoded output should be

rs3 rs4 rs5 rs6
1   0   0   0
1   0   1   0
0   0   0   0
1   0   1   0
0   0   0   0
0   2   0   1
0   2   NA  1
0   2   1   1
NA  1   1   1

I have 70000SNPs that need to check and recode. How to use the for loop or other method to do that in R?

Answer 1

Here's another possible (vectorized) solution

indx <- colSums(d == 2, na.rm = TRUE) < 3 # Select columns by condition
d[indx][d[indx] == 2] <- 1 # Inset 1 when the subset by condition equals 2
d
#   rs3 rs4 rs5 rs6
# 1   1   0   0   0
# 2   1   0   1   0
# 3   0   0   0   0
# 4   1   0   1   0
# 5   0   0   0   0
# 6   0   2   0   1
# 7   0   2  NA   1
# 8   0   2   1   1
# 9  NA   1   1   1

Answer 2

We can try

 d[] <- lapply(d, function(x) 
    if(sum(x==2, na.rm=TRUE) < 3) replace(x, x==2, 1) else x)
d
#   rs3 rs4 rs5 rs6
#1   1   0   0   0
#2   1   0   1   0
#3   0   0   0   0
#4   1   0   1   0
#5   0   0   0   0
#6   0   2   0   1
#7   0   2  NA   1
#8   0   2   1   1
#9  NA   1   1   1

Or the same methodology can be used in dplyr

library(dplyr)
d %>%
    mutate_each(funs(if(sum(.==2, na.rm=TRUE) <3) 
                replace(., .==2, 1) else .))