Question

I’m new to R, but I have a microarray dataset. I have an NA value at one location, and I try to use impute.knn() from library(impute), but after running the matrix through the impute.knn() function, it seems to turn into another structure. Can anyone explain what’s going on?

> m <- as.matrix(m)

# checking structure
> str(m)
  num [1:22283, 1:20] 1942.1 40.1 72.1 4693.6 35.9 ...
  - attr(*, "dimnames")=List of 2
  ..$ : chr [1:22283] "1007_s_at" "1053_at" "117_at" "121_at" ...
  ..$ : chr [1:20] "GSM146778-Normal" "GSM146780-Normal" "GSM146782-Normal"

# checking missing value of cell
> m[x,y] 
[1] NA

# impute missing value
> m.i <- impute.knn(m, 6)

# check value of imputed value
> m.i[x,y]
Error in r.imp["206054_at", "GSM146784-Normal"] : incorrect number of dimensions

# check structure of imputed matrix
> str(m.i)
List of 3
$ data     : num [1:22283, 1:20] 1942.1 40.1 72.1 4693.6 35.9 ...
..- attr(*, "dimnames")=List of 2
.. ..$ : chr [1:22283] "1007_s_at" "1053_at" "117_at" "121_at" ...
.. ..$ : chr [1:20] "GSM146778-Normal" "GSM146780-Normal" "GSM146782-Normal" "GSM146784-Normal" ...
$ rng.seed : num 3.62e+08
$ rng.state: int [1:626] 403 50 1992223309 -108730617 1600482030 698744776 

Answer 1

It looks like the impute.knn function returns a list which contains the matrix and other data regarding the imputation. Try using:

m.i$data[x,y]