Phylogenetic coding: “Up to now, there is no common agreement to either code characters as complex or simple.” – researchers

A friend calls attention to this interesting new paper:

When Homoplasy is not Homoplasy: Dissecting Trait Evolution by Contrasting Composite and Reductive Coding

Alejandro Torres-Montúfar  Thomas Borsch  Helga Ochoterena

Syst Biol syx053. DOI: https://doi.org/10.1093/sysbio/syx053

Published: 22 June 2017  Article history Received: 03 February 2017

Revision Received: 13 April 2017 Revision Received: 18 May 2017

Accepted: 24 May 2017

Abstract: The conceptualization and coding of characters is a difficult issue in phylogenetic systematics, no matter which inference method is used when reconstructing phylogenetic trees or if the characters are just mapped onto a specific tree. Complex characters are groups of features that can be divided into simpler hierarchical characters (reductive coding), although the implied hierarchical relational information may change depending on the type of coding (composite vs reductive). Up to now, there is no common agreement to either code characters as complex or simple. Phylogeneticists have discussed which coding method is best, but have not incorporated the heuristic process of reciprocal illumination to evaluate the coding. Composite coding allows to test 1) if several characters were linked resulting in a structure described as a complex character or trait, or 2) if independently evolving characters resulted in the configuration incorrectly interpreted as a complex character. We propose that complex characters or character states should be decomposed iteratively into simpler characters when the original homology hypothesis is not corroborated by a phylogenetic analysis, and the character or character state is retrieved as homoplastic. We tested this approach using the case of fruit types within subfamily Cinchonoideae (Rubiaceae). The iterative reductive coding of characters associated to drupes allowed us to unthread fruit evolution within Cinchonoideae. Our results show that drupes and berries are not homologous. As a consequence, a more precise ontology for the Cinchonoideae drupes is required.

 Character evolution, Complex characters, Simple characters, Rubiaceae, Fruit evolution, Drupes, Homoplasy  [colour emphasis added]

x