A multi‐platform package for the analysis of intra‐ and interspecific trait evolution

Abstract

1. Evolutionary forces affect the distribution of phenotypes both within and among species. Yet, at the macro-evolutionary scale, the evolution of intraspecific variance is rarely considered. Here, we present an r and a BEAST 2 implementation that extends the JIVE (Joint inter-and Intraspecific Variance Evolution) model aimed at the analysis of continuous trait evolution at both inter-and intraspecific level. 2. Using a hierarchical Bayesian approach, we implemented a range of models for continuous trait evolution that operate independently on species means and variances along a phylogeny. The package uses Markov chain Monte Carlo for the inference of parameters and the evaluation of model fit. JIVE is available in the bite (Bayesian Integrative models of Trait Evolution) r package, as well as in BEAST 2. The two implementations offer the same continuous trait evolutionary models, but differ in their use and types of analyses. The r implementation allows for faster analyses by taking the phylogeny as data, while providing graphical and statistical functions as part of tools for model comparison, result parsing and summary, and plotting. In the BEAST 2 implementation, the species tree is a parameter, and both its topology and divergence times are jointly estimated with trait model parameters. 3. The bite package and the BEAST 2 implementation introduce new frameworks within comparative phylogenetics that explicitly model intraspecific variance. These tools allow users to tackle long-standing questions in evolutionary biology, such as the identification of key evolutionary processes determining niche conservatism, niche partitioning, and life-history strategies.

Novel Integrative Modeling of Molecules and Morphology across Evolutionary Timescales

“…To further confirm the correctness of our implementation, we carried out an extensive well-calibrated validation study (e.g., Zhang et al ., Gaboriau et al . and Zhang, Drummond & Mendes [11, 61, 62]). This tests the implementation and illustrates statistical power and parameter identifiability.…”

On the Effect of Asymmetrical Trait Inheritance on Models of Trait Evolution

“…Evolutionary models for discrete traits already account for nonidentical inheritance at speciation, such as the “Dispersal-Extinction-Cladogenesis” model ( Ree et al 2005 ; Ree and Smith 2008 ), and the “Cladogenetic Stage change Speciation and Extinction” (ClaSSE) model ( Goldberg and Igić 2012 ). Analogous models for continuous traits could account for intra and interspecific variance ( Kostikova et al 2016 ; Gaboriau et al 2020 ) and directly treat individuals, rather than species, as the unit of phenotypic evolution. This would require trait measurements per species from multiple individuals, which will then allow us to infer the link between micro and macroevolutionary processes, formalized in our simulation framework.…”

Fast Bayesian inference of phylogenies from multiple continuous characters

“…Alvarez Carretero et al, 2019;May and Moore, 2020;Gaboriau et al, 2020). It was not until recently that careful simulation studies investigated the use of phylogenetic BM (Parins-Fukuchi, 2018b;Varón-González et al, 2020) and implemented statistical tools with the purpose of placing fossils and inferring phylogenies (Parins-Fukuchi, 2018a;…”

“…This strategy has been long recognized as an improvement to molecular substitution models (Sullivan and Swofford, 1997), and in our and similar cases might prevent the grouping of lineages that have undergone large amounts of phenotypic change. Local morphological clock models applied to evolutionary rates and adaptive optima regimes (in the case of OU models) could further accommodate ecologically relevant traits evolving under selection (Eastman et al, 2011;Uyeda and Harmon, 2014;Gaboriau et al, 2020). Accounting for rampant gene tree discordance due to incomplete lineage sorting and introgression may also prove necessary if it is shown that population-level processes cannot be merely buffered out through an additional variance term, but by being carefully modeled instead .…”

“…Over time we expect this last reason to become less of a hindrance and more of a motivation for method development in this area. Recent work suggests, however, there is an immediate demand for methods capable of integrating multidimensional data (e.g., Silvestro et al, 2018;Cascini et al, 2019;Koch and Thompson, 2020), as well as work in progress to meet those demands (e.g.,Álvarez Carretero et al, 2019;May and Moore, 2020;Gaboriau et al, 2020;Ogilvie et al, 2021).…”