Integration of Anatomy Ontologies and Evo-Devo Using Structured Markov Models Suggests a New Framework for Modeling Discrete Phenotypic Traits

Abstract

Modeling discrete phenotypic traits for either ancestral character state reconstruction or morphology-based phylogenetic inference suffers from ambiguities of character coding, homology assessment, dependencies, and selection of adequate models. These drawbacks occur because trait evolution is driven by two key processes—hierarchical and hidden—which are not accommodated simultaneously by the available phylogenetic methods. The hierarchical process refers to the dependencies between anatomical body parts, while the hidden process refers to the evolution of gene regulatory networks (GRNs) underlying trait development. Herein, I demonstrate that these processes can be efficiently modeled using structured Markov models (SMM) equipped with hidden states, which resolves the majority of the problems associated with discrete traits. Integration of SMM with anatomy ontologies can adequately incorporate the hierarchical dependencies, while the use of the hidden states accommodates hidden evolution of GRNs and substitution rate heterogeneity. I assess the new models using simulations and theoretical synthesis. The new approach solves the long-standing “tail color problem,” in which the trait is scored for species with tails of different colors or no tails. It also presents a previously unknown issue called the “two-scientist paradox,” in which the nature of coding the trait and the hidden processes driving the trait’s evolution are confounded; failing to account for the hidden process may result in a bias, which can be avoided by using hidden state models. All this provides a clear guideline for coding traits into characters. This article gives practical examples of using the new framework for phylogenetic inference and comparative analysis.

Inferring the Total-Evidence Timescale of Marattialean Fern Evolution in the Face of Model Sensitivity

“…3, 4) is a manifestation of the nonidentifiability of this model component: differences in time (node heights) can be compensated for by transition model-the F81+Γ mixture model-is also the 1120 most parameter-rich. This result may indicate that our 1121 morphological dataset could support even more com-1122 plex models, for example, ones that allow the process of evolution to vary among branches (Beaulieu et al 2013;Goloboff et al 2019), or that accommodate correlated evolution among characters (Pagel and Meade 2006;Meyer et al 2019) and other complex dependencies (Maddison 1993;Tarasov 2019). However, such models generally require increasing the state-space of the characters, which complicates the calculations used for correcting for variable-only characters, so more work is necessary before they should be used in total-evidence dating analyses.…”

Interaction among ploidy, breeding system and lineage diversification

“…In order to compare fits of ploidy‐only vs breeding system‐only vs combined trait models, we use the technique of ‘lumping’ states together (Tarasov, ). We use the state space of the ID / CD / CP model but constrain the rate parameters to mimic the behavior of the single‐trait models.…”

“…We use the state space of the ID / CD / CP model but constrain the rate parameters to mimic the behavior of the single‐trait models. Lumping states requires that the transition rates from the lumped state to the singular state be equal (Tarasov, ). First, we lump together ID and CD to form the diploid state, mimicking the D / P model (M26).…”

Mismatch of the morphology model is mostly unproblematic in total-evidence dating: insights from an extensive simulation study

“…Observed character states in morphology are believed to be difficult to delimit in an objective way, and transitions between them might not adequately reflect the underlying biochemical and developmental processes (Tarasov 2019). Furthermore, morphological characters are generally assumed to have a low clocklikeness (Beck and Lee 2014;Lee 2016;).…”

“…Alekseyenko et al (2008) proposed a combination of an Mk-like model and a binary stochastic Dollo-model based on the Poisson process, thus allowing for two types of transitions within a single character; this might be an adequate depiction of multi-state characters that include an "absent" state and several modifications of the "present" state, as in the famous example of a tail with different colours. Tarasov (2019) went even further by proposing hidden-state Markov models that reflect the mismatch between our character concepts and the underlying evolutionary process. All these models promise to lead to more robust phylogenetic inference from morphological data and improve our understanding of morphological evolution; however, it remains unclear how much model realism is in fact needed to obtain accurate divergence-time estimates in the TED framework.…”