CODEWITHSUNDEEP
rrandom
colin
colin

Reputation: 2666

Randomly sample columns of a matrix with replacement, then sample elements of a vector in the same order - R

I am performing a bootstrap analysis of community similarity values. I have a matrix of species counts, where the columns represent unique samples, and the rows unique species. Here is an example:

#generate a matrix with 20 columns, 30 rows, random values
set.seed(69) #for reproducibility.
otu <-matrix(rpois(20*30, lambda = 2), ncol=20)

I have two vectors that are environmental covariates associated with the unique samples. Therefore, each of these vectors is length 20.

v1 <- rnorm(20)
v2 <- rnorm(20)

I want to create a new species observation matrix that randomly samples the columns of the otu matrix, with replacement. I can do this fairly easily.

#randomnly sample the columns of the otu matrix with replacement.
otu.boot <- otu[,sample(ncol(otu),size=ncol(otu),replace=TRUE)]

My question is, how can I sample vectors v1 and v2 in the same order as the randomly sampled with replacement matrix?

Upvotes: 0

Views: 378

Answers (1)

colin
colin

Reputation: 2666

Following @lmo's suggestion in the comments:

#establish a sample order with replacement.
 myOrder <- sample(ncol(otu), replace = T)

#go ahead and submsample both the otu matrix and vectors in that order.
otu.boot <- otu[,myOrder]
 v1.boot <-  v1[myOrder]
 v2.boot <-  v2[myOrder]

Upvotes: 0

Related Questions