The ancestral function prediction and SF, CF and NF classification results on two synthetic cases (A) and (B). The panel on the left shows synthetic data according to SF, CF and NF rules, where we synthesize the expression levels at internal nodes. The figures on the right shows the prediction results by the NBP algorithm, where "Mean" denotes the mean value of the Gaussian mixture component with the largest mixing weight denoted by the "Weight". We also show the classification probabilities of SF, CF and NF at internal nodes according to Eq. (8), where only the class with the highest probability is shown.