Table 2.
Comparisons of coalescent-based ASTRAL species phylogenies with and without collapsing poorly supported bipartitions in replicate gene trees.
| Data set | Program | Without collapsing branches | With collapsing branches | ||
|---|---|---|---|---|---|
| Conflicting bipartitions (%)a | Highly conflicting bipartitions (%)b | Conflicting bipartitions (%)a | Highly conflicting bipartitions (%)b | ||
| Animal: Bees | IQ-TREE | 22.46 | 1.87 | 18.18 | 1.60 |
| Animal: Birds | IQ-TREE | 6.09 | 0 | 6.60 | 0 |
| Animal: Butterflies | IQ-TREE | 7.35 | 0.49 | 5.39 | 0.74 |
| Animal: Lizards | IQ-TREE | 7.69 | 0 | 0 | 0 |
| Animal: Marine fishes | IQ-TREE | 7.69 | 0 | 13.68 | 0 |
| Animal: Rodents | IQ-TREE | 0 | 0 | 0 | 0 |
| Plant: Cardueae | IQ-TREE | 12.20 | 0 | 12.20 | 0 |
| Plant: Caryophyllales | IQ-TREE | 0 | 0 | 0 | 0 |
| Plant: Green plants | IQ-TREE | 5.70 | 0.38 | 5.53 | 0.55 |
| Plant: Jaltomata | IQ-TREE | 8.33 | 0 | 0 | 0 |
| Plant: Protea | IQ-TREE | 14.52 | 0 | 3.23 | 0 |
| Fungi: Aspergillaceae | IQ-TREE | 0 | 0 | 0 | 0 |
| Fungi: Budding yeasts | IQ-TREE | 0.29 | 0 | 0.29 | 0 |
| Fungi: Hanseniaspora | IQ-TREE | 0 | 0 | 0 | 0 |
| Fungi: Rhizoplaca | IQ-TREE | 3.57 | 0 | 0 | 0 |
| Animal: Bees | RAxML-NG | 27.27 | 1.87 | 8.02 | 0 |
| Animal: Birds | RAxML-NG | 4.06 | 0 | 0 | 0 |
| Animal: Butterflies | RAxML-NG | 6.86 | 0.49 | 6.86 | 0 |
| Animal: Lizards | RAxML-NG | 19.23 | 0 | 0 | 0 |
| Animal: Marine fishes | RAxML-NG | 3.42 | 0 | 5.98 | 0 |
| Animal: Rodents | RAxML-NG | 0 | 0 | 0 | 0 |
| Plant: Cardueae | RAxML-NG | 12.20 | 0 | 3.66 | 0 |
| Plant: Caryophyllales | RAxML-NG | 0 | 0 | 0 | 0 |
| Plant: Green plants | RAxML-NG | 0.26 | 0 | 1.02 | 0.09 |
| Plant: Jaltomata | RAxML-NG | 8.33 | 0 | 0 | 0 |
| Plant: Protea | RAxML-NG | 14.52 | 0 | 8.06 | 0 |
| Fungi: Aspergillaceae | RAxML-NG | 0 | 0 | 0 | 0 |
| Fungi: Budding yeasts | RAxML-NG | 0.29 | 0 | 0.59 | 0 |
| Fungi: Hanseniaspora | RAxML-NG | 0 | 0 | 0 | 0 |
| Fungi: Rhizoplaca | RAxML-NG | 3.57 | 0 | 0 | 0 |
Since running all 19,414 gene alignments from 15 phylogenomic data sets was computationally intractable, we sampled the first 100 genes from each data set. For each data set, coalescent-based ASTRAL trees were reconstructed from the Run1 and Run2 sets of 100 individual gene trees without and with collapsing branches with low bootstrap support values (≤50%); both Run1 and Run2 used identical settings, including substitution model, random starting seed number, number of threads of execution, number of independent tree searches, number of bootstrap replicates, and ML program.
aPercentage of conflicting bipartitions between coalescent-based ASTRAL species phylogenies inferred using Run1 and Run2 gene trees.
bPercentage of highly conflicting bipartitions (LPP ≥ 90%) between coalescent-based ASTRAL species phylogenies inferred using Run1 and Run2 gene trees.