Potts models

Potts models, also called Markov Random Fields, are pairwise probability models assembled from MSAs. Unlike PSSMs, which only capture position-specific probabilities, Potts models capture second-order pairwise probabilities as well.