[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Prolacertiformes and Protorosauria

David Peters wrote:

re: the more robust tree question: The tree that has the fewest mpts and survives the most decay tests is the more robust tree. Is it not? In other words, the tree in which sister taxa share the most synapomorphies is the more robust tree, yes?

The number of MPTs is not actually a particularly strong measure of the robustness of the tree. Surviving decay tests is a better indicator, but it still doesn't actually tell you if your dataset is stronger than another dataset - it indicates whether or not you are confident reporting the MPT in question, relative to the next-most parsimonious trees. Decay tests are mostly a way of approaching the issue that biological systems are not actually parsimonious. If you can relax the parsimony requirement a great deal, and still recover the same topology, then the tree is at least robust to that particular assumption.

However, those indices do not answer the question of preference of topology *between* datasets - when two different matrices produce different topologies, then decay indices alone are not particular informative. If one of the matrices produces many MPTs, with sensitivity to decay tests, then that particular analysis is hypothesizing a lack of resolution - such a result essentially indicates that we don't have a good idea of the answer. If the competing matrix finds few MPTs, containing clades robust to decay tests, then we have the opposite result: that analysis would support the conclusion that we have high confidence in the relationships. The problem is that, without additional information, we have no way of judging which hypothesis is correct: a lack of knowledge result is no less plausible than a presence of knowledge result.



Michael Habib, M.S. PhD. Candidate Center for Functional Anatomy and Evolution Johns Hopkins School of Medicine 1830 E. Monument Street Baltimore, MD 21205 (443) 280-0181 habib@jhmi.edu