Nested Clades, The Consistency Index, and Affirming the Consequent

[EricMH (Eric Holloway)]

@paleomalacologist

Klassen et al. 1991 (Consistency Indices and Random Data on JSTOR) argued that finding that the CI was above that expected for a random data set of similar size was a useful indicator that there is phylogenetic signal in the data set.

That is exactly what I did above. I showed DAGs produce CI way above the chance cuttoff.

@paleomalacologist

Instead, analyses focus on the use of analyses such as maximum parsimony, maximum likelihood, or Bayesian analyses to determine what are the best trees.

Did you see my response where I did this? Here it is again:

Nested Clades, The Consistency Index, and Affirming the Consequent

Ha! Exactly what I did at first, at the gene level, and showed DAGs produce distinct trees just like genealogical processes do. The feedback was I needed to do this at the nucleotide level with official tools. Now, I’m not very familiar with the PAUP software, but I think this addresses what you want. I create 100 random trees, run a likelihood test and sort the trees. If the clade structure is the same, then the sorted range should be pretty small, and visa versa. There is a large DAG dat…

I ran a likelihood analysis on my DAG dataset and compared to a real world dataset, and it was indistinguishable.

Anyways, can you recommend a specific process/metric I can use that is supposed to reliably indicate an evolutionary process created the dataset?

I had thought this CI thing was a really well established result from the Klassen 1991 paper that analyzed 70+ CI results. But now you are telling me that CI metric is meaningless. So, what do we make of the Klassen 1991 paper and their analysis of 70+ CI results? Is that also meaningless?

The field of evolutionary science is really confusing!!!