Fast and accurate estimation of species-specific diversification rates using data augmentation

Maliet, Odile; Morlon, Hélène

doi:10.1101/2020.11.03.365155

Cited by 6 publications

(3 citation statements)

References 56 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using the most comprehensive squamate phylogeny available to date ( 50 ), we estimated speciation rates using both a model-based [

CLaDS; ( 8 , 51 )] and a semiparametric approach [

DR ; ( 6 )]. These estimates of speciation rate are highly correlated both across all squamates ( n = 9,755, r = 0.87) and across our focal taxa ( n = 59, r = 0.87; SI Appendix , Fig.…”

Section: Resultsmentioning

confidence: 99%

“…Although the metric weights recent splitting events more, it estimates tip speciation rates with reasonable accuracy, even in the presence of extensive stochastic variation in branch lengths ( 5 ). For a model-based approach, we used CLaDS (cladogenetic diversification rate shift model; 51 ) to estimate speciation rates (

CLaDS ). CLaDS applies a Bayesian approach to inferring speciation rates along a phylogeny and assumes that rates change after every speciation event.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

No link between population isolation and speciation rate in squamate reptiles

Singhal

Colli

Grundler

et al. 2022

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

Rates of species formation vary widely across the tree of life and contribute to massive disparities in species richness among clades. This variation can emerge from differences in metapopulation-level processes that affect the rates at which lineages diverge, persist, and evolve reproductive barriers and ecological differentiation. For example, populations that evolve reproductive barriers quickly should form new species at faster rates than populations that acquire reproductive barriers more slowly. This expectation implicitly links microevolutionary processes (the evolution of populations) and macroevolutionary patterns (the profound disparity in speciation rate across taxa). Here, leveraging extensive field sampling from the Neotropical Cerrado biome in a biogeographically controlled natural experiment, we test the role of an important microevolutionary process—the propensity for population isolation—as a control on speciation rate in lizards and snakes. By quantifying population genomic structure across a set of codistributed taxa with extensive and phylogenetically independent variation in speciation rate, we show that broad-scale patterns of species formation are decoupled from demographic and genetic processes that promote the formation of population isolates. Population isolation is likely a critical stage of speciation for many taxa, but our results suggest that interspecific variability in the propensity for isolation has little influence on speciation rates. These results suggest that other stages of speciation—including the rate at which reproductive barriers evolve and the extent to which newly formed populations persist—are likely to play a larger role than population isolation in controlling speciation rate variation in squamates.

show abstract

“…Using the most comprehensive squamate phylogeny available to date ( 50 ), we estimated speciation rates using both a model-based [

CLaDS; ( 8 , 51 )] and a semiparametric approach [

DR ; ( 6 )]. These estimates of speciation rate are highly correlated both across all squamates ( n = 9,755, r = 0.87) and across our focal taxa ( n = 59, r = 0.87; SI Appendix , Fig.…”

Section: Resultsmentioning

confidence: 99%

CLaDS ). CLaDS applies a Bayesian approach to inferring speciation rates along a phylogeny and assumes that rates change after every speciation event.…”

Section: Methodsmentioning

confidence: 99%

No link between population isolation and speciation rate in squamate reptiles

Singhal

Colli

Grundler

et al. 2022

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

show abstract

“…As a result, there are several models of diversification in the literature which behavior has been studied with simulations but that lack a proper inference machinery (McPeek 2008;Aristide and Morlon 2019;Hagen et al 2021). Methods based on Expectation Maximization (EM) algorithms (Dempster et al 1977;Richter et al 2020), data augmentation (Maliet and Morlon 2022), or composite likelihoods (Lindsay 1988;Varin et al 2021)) can overcome some of these limitations (Raynal 2019), yet they still rely on likelihood formulae.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning from Phylogenies for Diversification Analyses

Lambert

Voznica

Morlon

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Birth-death models are widely used in combination with species phylogenies to study past diversification dynamics. Current inference approaches typically rely on likelihood-based methods. These methods are not generalizable, as a new likelihood formula must be established each time a new model is proposed; for some models such formula is not even tractable. Deep learning can bring solutions in such situations, as deep neural networks can be trained to learn the relation between simulations and parameter values as a regression problem. In this paper, we adapt a recently developed deep learning method from pathogen phylodynamics to the case of diversification inference, and we extend its applicability to the case of the inference of state-dependent diversification models from phylogenies associated with trait data. We demonstrate the accuracy and time efficiency of the approach for the time constant homogeneous birth-death model and the Binary-State Speciation and Extinction model. Finally, we illustrate the use of the proposed inference machinery by reanalyzing a phylogeny of primates and their associated ecological role as seed dispersers. Deep learning inference provides at least the same accuracy as likelihood-based inference while being faster by several orders of magnitude, offering a promising new inference approach for deployment of future models in the field.

show abstract

Universal probabilistic programming offers a powerful approach to statistical phylogenetics

Ronquist

Kudlicka

Senderov

et al. 2020

Preprint

View full text Add to dashboard Cite

Statistical phylogenetic analysis currently relies on complex, dedicated software packages, making it difficult for evolutionary biologists to explore new models and inference strategies. Recent years have seen more generic solutions based on probabilistic graphical models, but this formalism can only partly express phylogenetic problems. Here we show that universal probabilistic programming languages (PPLs) solve the model expression problem, while still supporting automated generation of efficient inference algorithms. To illustrate the power of the approach, we use it to generate sequential Monte Carlo (SMC) algorithms for recent biological diversification models that have been difficult to tackle using traditional approaches. This is the first time that SMC algorithms have been available for these models, and the first time it has been possible to compare them using model testing. Leveraging these advances, we re-examine previous claims about the performance of the models. Our work opens up several related problem domains to PPL approaches, and shows that few hurdles remain before PPLs can be effectively applied to the full range of phylogenetic models.In statistical phylogenetics, we are interested in learn-1 ing the parameters of models where evolutionary trees-2 phylogenies-play an important part. Such analyses have a 3 surprisingly wide range of applications across the life sci-4 ences 1,2,3 . In fact, the research front in many disciplines is 5 partly defined today by our ability to learn the parameters 6 of realistic phylogenetic models. 7 Statistical problems are often analyzed using generic 8 modeling and inference tools. Not so in phylogenetics, 9 where empiricists are largely dependent on dedicated soft-10 ware developed by small teams of computational biolo-11 gists 3 . Even though these software packages have become 12 increasingly flexible in recent years, empiricists are still 13 limited to a large extent by predefined model spaces and 14 inference strategies. Venturing outside these boundaries 15 typically requires the help of skilled programmers and in-16 ference experts. 17 If it were possible to specify arbitrary phylogenetic mod-18 els in an easy and intuitive way, and then automatically 19 learn the latent variables (the unknown parameters) in them, 20 the full creativity of the research community could be un-21 leashed, significantly accelerating progress. There are two 22 major hurdles standing in the way of such a vision. First, we 23 must find a formalism (a language) that can express phyloge-24 netic models in all their complexity, while still being easy to 25 learn for empiricists (the model expression problem). Sec-26 ond, we need to be able to generate computationally efficient 27 inference algorithms from such model descriptions, draw-28 ing from the full range of techniques available today (the 29 automated inference problem).30In recent years, there has been significant progress to-31 wards solving the model expression problem by adopting 32 the framework of probabilistic graphi...

show abstract

Fast and accurate estimation of species-specific diversification rates using data augmentation

Cited by 6 publications

References 56 publications

No link between population isolation and speciation rate in squamate reptiles

No link between population isolation and speciation rate in squamate reptiles

Deep Learning from Phylogenies for Diversification Analyses

Universal probabilistic programming offers a powerful approach to statistical phylogenetics

Contact Info

Product

Resources

About