Additional file 4: of Phylogenomic analysis of Apoidea sheds new light on the sister group of bees

Figure 2. Results from the Four-cluster Likelihood Mapping showing the support for the possible relationship of Ammoplanina (two species), Psenini (four species) + Odontosphecini (one species), Anthophila (42 species) and all remaining species including outgroup species to each other (58050 quartets). Original amino acid supermatrix (94,869 amino acid sites) based on a protein domain-based partitioning scheme and analyzed with partition-specific substitution models (a), permutation scheme I with supermatrix as in a, but amino acids are permuted within partitions while retaining the specific distribution of missing data (b), permutation scheme II with supermatrix as in b, but replacing amino acids in each partition with randomly selected amino acids, using amino acid frequencies as given by the LG substitution matrix, while retaining the specific distribution of missing data (c) and permutation scheme III supermatrix as given in c, but with missing data being randomly permuted (d). Figure 3. Results from the Four-cluster Likelihood Mapping showing the support for the possible relationship of Mellininae (one species), Sphecidae (19 species), Crabroninae (39 species) + Dinetinae (one species), and all remaining species including bees and outgroup species to each other (103320 quartets). Original amino acid supermatrix (94,869 amino acid sites) based on a protein domain-based partitioning scheme and analyzed with partition-specific substitution models (a), permutation scheme I with supermatrix as in a, but amino acids are permuted within partitions while retaining the specific distribution of missing data (b), permutation scheme II with supermatrix as in b, but replacing amino acids in each partition with randomly selected amino acids, using amino acid frequencies as given by the LG substitution matrix, while retaining the specific distribution of missing data (c) and permutation scheme III supermatrix as given in c, but with missing data being randomly permuted (d). Figure 4. Results from the Four-cluster Likelihood Mapping showing the support for the possible relationship of Ammoplanina (two species), Psenini (four species) + Odontosphecini (one species), Anthophila (42 species) and all remaining species including outgroup species to each other (58050 quartets). Original nucleotide supermatrix (284.607 nucleotide sites) partitioned based on applying a combination of protein domain – and codon-based partitioning scheme by modelling the 1st, 2nd and 3rd codon position separately. Each partition was analyzed with partition-specific model parameters under the nucleotide substitution model GTR+G (a), permutation scheme I with supermatrix as in a, but nucleotides are permuted within partitions while retaining the specific distribution of missing data (b), permutation scheme II with supermatrix as in b, but replacing nucleotides in each partition with randomly selected nucleotides, while retaining the specific distribution of missing data (c) and permutation scheme III supermatrix as given in c, but with missing data being randomly permuted (d). Figure 5. Results from the Four-cluster Likelihood Mapping showing the support for the possible relationship of Mellininae (one species), Sphecidae (19 species), Crabroninae (39 species) + Dinetinae (one species), and all remaining species including bees and outgroup species to each other (103320 quartets). Original nucleotide supermatrix (284.607 nucleotide sites) partitioned based on applying a combination of protein domain – and codon-based partitioning scheme by modelling the 1st, 2nd and 3rd codon position separately. Each partition was analyzed with partition-specific model parameters under the nucleotide substitution model GTR+G (a), permutation scheme I with supermatrix as in a, but nucleotides are permuted within partitions while retaining the specific distribution of missing data (b), permutation scheme II with supermatrix as in b, but replacing nucleotides in each partition with randomly selected nucleotides, while retaining the specific distribution of missing data (c) and permutation scheme III supermatrix as given in c, but with missing data being randomly permuted (d). (DOC 1269 kb