Table 1 Van Lishout’s MaxT

(1) Compute the test-statistics for all pairs, but only store the n highest tests values. The result is a Real data

vector where T _0,1≥T _0,2≥…≥T _0,n .

(2) Initialise a vector p of size n with 1’s.

(3) Perform the following operations for i=1,…,B:

(a) Generate a random permutation of the trait column.

(b) Compute T _i,1,…,T _i,n and store them in a Permutation _i vector.

(c) Compute the maximum M _i of the test-statistics values T _i,n+1,…,T _i,m .

(d) Replace T _i,n by M _i if T _i,n <M _i .

(e) Force the monotonicity of the Permutation _i vector: for j=n−1,…,1 replace T _i,j by T _i,j+1 if T _i,j <T _i,j+1.

(f) For each j=1,…,n, if T _i,j ≥T _0,j increment p _j by one.

(4) Divide all values of vector p by B+1 to obtain the p-values vector. Force monotonicity: for j=1,…,n−1,

replace p _j+1 by p _j if p _j+1<p _j .