This is how a phylogenetic tree comes out after obtaining an alignment from one of these DAGs(containing 30 “taxa”):
I just obtained a new DAG from the link he gives to his data generator tool.
I first sent the “taxa”(which comes in fasta format) from the DAG to a multiple sequence alignment tool (MAFFT) using default settings.
Btw the alignment consists mostly of gaps as the sequences from different taxa have wildly different lengths.
For example this is one of his taxa:
Then sent the alignment to a ML program again using default settings, but with bootstrap 100 turned on.
Had to force it into dendrogram mode to see the bootstrap values:
Most of them are pretty crap
Regardless, none of this makes much sense because I’m still not sure how Holloway generates his data. It appears like his terminal taxa basically consists of anything output by his DAG generator, be they ancestors or descendants. I’m not sure.
I suspect that, to the extend there is any tree-like structure in the data, it’s because his DAG actually happened to implement a tree-like process, so if and when Holloway claims to measure any noteworthy CI, it’s because his DAG really has coincidentally implemented something like common descent.