The number of clusters K for each disease was chosen by minimizing the Bayesian information criterion (BIC) complexity penalty: −2L̂k + KJ log (N), where L̂k indicates the sum of ℒK values over the N input samples, computed using K clusters. In order to avoid local minima, the EM algorithm was run 25 times for each value of K ∈ [2,8] with randomized starting points and the best model was retained.
