Influence of removing substantial proportions of fast-evolving sites in the amino acid data set on the resolution of phylogenetic relationships. (A) Best ML tree inferred from the data subset lacking 70% of the original information. (B) Best ML tree inferred from the data subset lacking 90% of the original information. The nodes that received 100% bootstrap support in ML and MP analyses are denoted by asterisks. For the other nodes, only the bootstrap values ≥ 50% are indicated; the values obtained in ML and MP analyses appear on the left and right, respectively.