This is probably more of a CS question than biology:

Given a set of m protein sequences I want to select n candidates out of this set (n is a given number), which maximize the diversity.

I would probably start with a distance matrix (made by clustalo). Now I want to choose my n candidates in this way, that the total sum of distance of each candidate to every other candidate is maximized.

The goal is to get a subset of the m protein sequences, which is still more or less representative in terms of the diversity.

What approach would you suggest?