IT benchmark with column likelihood

Nikolai's IT benchmark.
Finish consensus tree compilation w astral
finish column likelihood infererence
compile master dataframe
perform discretized double integral for MI over columns

other metrics with MI for likelihood?
column diversity? distance from typical distribution of characters / bg frequencies?


