Question

I’m a newbie to R and I have a dataframe called pedM with 203 rows and 19 columns and I wanted to code a line to delete then columns that returned mostly a value of 0 in the dataframe (say, 98% of the column is 0).

I’m using the dataframe to output a heatmap which returns an image based on gene expression, and I want to get rid of the lines that have outliers that don’t really contribute to the data.

I imagine the line of code I’m looking for is fairly simple, but I can’t wrap my head around the correct way to do it. Just to reiterate, looking for a line that would basically remove a column from a dataframe that has 98% 0’s or 2% nonzero’s, whichever way is easier.

Thanks in advance.

Vivek

Answer 1

I like to do this in a few steps for clarity. First define a function that checks a vector for 98% zeros and returns True or False:

myfun <- function(x, threshhold=0.98) {
  if(length(x[x==0])/length(x) > threshhold) {
    TRUE
  } else {
    FALSE
  }
}

Then create a boolean vector for each column using apply:

to.drop <- apply(pedM, 2, myfun)

Finally drop the columns you don’t want:

pedM <- pedM[, !to.drop]