Search code examples
rmatrixchemistry

Calculate diversity index (dissimilarity index) for a set of compounds in R

I want to calculate the diversity index for a given matrix.

I have a dataset matrix (xmatrix.RData), which is a 986 * 881 matrix, indicating 986 compounds and 881 fingerprints descriptors.

The formula to calculate the diversity index is explained in:
http://r.789695.n4.nabble.com/file/n4677766/Diversity_Index_Formula.pdf

Solution

  • I would do something like this:

    # this calulates you the dissimilarity matrix diss(i, j) in the paper
# look at help page of dist for different methods like euclidean, maximum...
diss<-as.matrix(dist(xmatrix, method="binary", diag=TRUE, upper=TRUE))
l<-nrow(xmatrix)

# the overall dissimilarity
sum(rowSums(diss))/(l*(l-1))

    hope this helps...