Table 2.
Data Set | Partitioning Scheme | Topological Constraint | Number of Partitions | Substitution Model | Number of Parameters | Ln L | AIC | ΔAIC |
---|---|---|---|---|---|---|---|---|
All genes | Unpartitioned (one partition) | None | 1 | GTR | 8 | −787,773 | 1,575,562 | 62,885 |
PartitionFinder (six partitions) | On one partition tree | 6 | GTR | 48 | −758,061 | 1,516,219 | 3,349 | |
Gene/codon-position (41 partitions) | On one partition tree | 41 | GTR | 328 | −756,379 | 1,513,414 | 737 | |
Gene/codon-position (41 partitions) | On six partition tree | 41 | GTR | 328 | −756,272 | 1,513,199 | 522 | |
PartitionFinder (six partitions) | On 41 partition tree | 6 | GTR | 48 | −758,010 | 1,516,116 | 3,439 | |
Gene/codon-position (41 partitions) | None | 41 | GTR | 328 | −756,010 | 1,512,677 | n/a | |
PartitionFinder (six partitions) | On one partition tree | 6 | GTR | 48 | −758,061 | 1,516,219 | 3,542 | |
Protein-coding genes | Unpartitioned (one partition) | None | 1 | GTR | 8 | −684,161 | 1,368,339 | 34,473 |
Gene/codon-position (39 partitions) | On 1 partition tree | 39 | GTR | 312 | −666,834 | 1,334,219 | 425 | |
PartitionFinder (5 partitions) | None | 5 | GTR | 40 | −668,480 | 1,337,039 | 3,173 | |
Gene/codon-position (39 partitions) | On five partition tree | 39 | GTR | 312 | −666,678 | 1,333,981 | 115 | |
PartitionFinder (five partitions) | On 39 partition tree | 5 | GTR | 40 | −668,523 | 1,337,127 | 3,261 | |
Gene/codon-position (39 partitions) | None | 39 | GTR | 312 | −666,621 | 1,333,866 | n/a | |
PartitionFinder (five partitions) | On one partition tree | 5 | GTR | 40 | −668,567 | 1,337,213 | 3,347 |
Note.—Trees were obtained under no partitioning, under the six- or five-partition schemes selected by PartitionFinder, and by the maximum number of partitions tested (partitioning by gene and codon position). Each of the resulting trees was then assessed for their likelihood under the alternative models. Note the comparatively small difference in likelihood (ΔAIC) under each partitioning scheme regardless of the model used in the tree search.