A Simple Statistical Method for Estimating Type-II (Cluster-Specific) Functional Divergence of Protein Sequences