Principled, practical, flexible, fast: a new approach to phylogenetic factor analysis

Hassler, Gabriel W.; Gallone, Brigida; Aristide, Leandro; Allen, William L.; Tolkoff, Max R.; Holbrook, Andrew J.; Baele, Guy; Lemey, Philippe; Suchard, Marc A.

doi:10.48550/arxiv.2107.01246

Cited by 3 publications

(7 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We performed PFA using the Julia package PhylogeneticFactorAnalysis.jl ver. 0.1.4 (Hassler et al 2021) which relies on a development version of BEAST (Suchard et al 2018) to be released with BEAST ver. 1.10.5.…”

Section: Trait Data and Phylogenetic Factor Analysismentioning

confidence: 99%

See 1 more Smart Citation

Chameleon biogeographic dispersal is associated with extreme life history strategies

et al. 2022

Self Cite

View full text Add to dashboard Cite

Understanding the role of traits in dispersal is necessary to improve our knowledge of historical biogeography, community assembly processes and predictions of species' future movements. Here we aimed to determine the relationship between three traits (coastal distribution, body size, position on the fast/slow life history continuum) and past dispersal probability on an evolutionary timescale in chameleons (Chamaeleonidae). Using species' distribution data we identified the nine most important biogeographic regions for all included chameleons (181/217 species). We compiled life history trait data and used phylogenetic factor analysis to infer independent body size and fast/slow life history trait axes. Finally, we tested whether traits and trait combinations related to biogeographic dispersal success in the past, using trait‐dependent biogeographic models. All three traits were associated with past biogeographical movements. Lineages having coastal distributions and those with large bodies had higher dispersal probabilities. Interestingly, chameleons with either a very fast or very slow life history were more successful dispersers than species with an intermediate strategy. Together, the three traits – coastal, large‐bodied and extreme life history – form a dispersal syndrome. Traits have played an important role in the biogeographic history of chameleons. While only fast traits have been linked to present‐day invasion success in reptiles, both extremes of the life history spectrum were likely advantageous for dispersal and establishment during past biogeographic movements. Fast‐living species may be less susceptible to stochastic extinction in the first phases of a colonization (due to rapid population growth), and slow‐living species may be less vulnerable to environmental stochasticity (due to low demographic variability). Our results call for broader analyses testing the general influence of life history strategy in biogeographic dispersal success, which would help explain species distribution patterns on Earth.

show abstract

Section: Trait Data and Phylogenetic Factor Analysismentioning

confidence: 99%

“…Q19 AQ: Graham and Fine 2008 provided in the list, but not cited in the text. Q20 Please update the reference 'Hassler et al 2021'. Q21 AQ: IUCN 2019.…”

Section: Author Contributionsmentioning

confidence: 99%

Chameleon biogeographic dispersal is associated with extreme life history strategies

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Second, though our method achieves the current best inference efficiency under the phylogenetic probit model, there is still room for improvement. One potential solution is to de-correlate some latent parameters by grouping them into independent factors using phylogenetic factor analysis (Tolkoff et al, 2018;Hassler et al, 2021). Also, we can consider a logistic or softmax function to map latent parameters to the probablity of a discrete trait.…”

Section: Discussionmentioning

confidence: 99%

Hamiltonian zigzag accelerates large-scale inference for conditional dependencies between complex biological traits

Zhang¹,

Nishimura²,

Trovão³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Inferring correlation among mixed-type biological traits while controlling for the evolutionary relationship among taxa is of great scientific interest yet remains computationally challenging. The recently developed phylogenetic multivariate probit model accommodates binary and continuous traits by assuming latent parameters underlying binary traits. The most expensive inference step is to sample the latent parameters from their conditional posterior that is a high dimensional truncated normal. The current best approach uses the bouncy particle sampler (BPS) optimized with a linear-order gradient evaluation method that employs a dynamic programming strategy on the directed acyclic structure of the phylogeny. Despite its significant improvement upon previous methods, with increasing sample sizes BPS encounters difficulty in exploring the parameter space and fails to provide reliable estimates for the across-trait partial correlation that describes the direct effects among traits. We develop a new inference scheme that highlights Zigzag Hamiltonian Monte Carlo (Zigzag-HMC), a variant of traditional HMC that uses Laplace momentum. Zigzag-HMC can utilize the same gradient evaluation method that speeds up BPS, yet it is much more efficient. We further improve the efficiency by jointly updating the latent parameters and correlation elements using a differential operator splitting technique. In an application exploring HIV-1 evolution that requires joint sampling from a 11,235-dimensional truncated normal and a 24-dimensional covariance matrix, our method yields a 5× speedup compared to BPS and makes it possible to estimate the direct effects among important viral mutations and virulence. We also extend the phylogenetic probit model to categorical traits for broader applicability, and demonstrate its use to study Aquilegia flower and pollinator co-evolution.

show abstract

“…, N with overall complexity O(N 2 PK 2 ). Hassler et al (2021) apply the likelihood calculation and data augmentation algorithms of Hassler et al (2020) to sample from X | Y, F, L, σ in O(N PK 3 ). As K is by design small, the cubic scaling in K is preferable to the quadratic scaling in N. Hassler et al (2021) also develop a novel HMC approach to efficiently sample directly from L | Y, F, σ without conditioning on the latent factors X that applies to latent factor models generally.…”

Section: 33mentioning

confidence: 99%

“…Hassler et al (2021) apply the likelihood calculation and data augmentation algorithms of Hassler et al (2020) to sample from X | Y, F, L, σ in O(N PK 3 ). As K is by design small, the cubic scaling in K is preferable to the quadratic scaling in N. Hassler et al (2021) also develop a novel HMC approach to efficiently sample directly from L | Y, F, σ without conditioning on the latent factors X that applies to latent factor models generally. Hassler et al (2021) show that one can calculate the gradient ∇ L logp(L | Y, F, σ ) required for HMC as a function of the full conditional mean and variance of each x i , but not the values of x i explicitly.…”

Section: 33mentioning

confidence: 99%

Data Integration in Bayesian Phylogenetics

Hassler

Magee

Zhang

et al. 2023

Annu. Rev. Stat. Appl.

View full text Add to dashboard Cite

Researchers studying the evolution of viral pathogens and other organisms increasingly encounter and use large and complex data sets from multiple different sources. Statistical research in Bayesian phylogenetics has risen to this challenge. Researchers use phylogenetics not only to reconstruct the evolutionary history of a group of organisms, but also to understand the processes that guide its evolution and spread through space and time. To this end, it is now the norm to integrate numerous sources of data. For example, epidemiologists studying the spread of a virus through a region incorporate data including genetic sequences (e.g., DNA), time, location (both continuous and discrete), and environmental covariates (e.g., social connectivity between regions) into a coherent statistical model. Evolutionary biologists routinely do the same with genetic sequences, location, time, fossil and modern phenotypes, and ecological covariates. These complex, hierarchical models readily accommodate both discrete and continuous data and have enormous combined discrete/continuous parameter spaces including, at a minimum, phylogenetic tree topologies and branch lengths. The increased size and complexity of these statistical models have spurred advances in computational methods to make them tractable. We discuss both the modeling and computational advances, as well as unsolved problems and areas of active research. Expected final online publication date for the Annual Review of Statistics and Its Application, Volume 10 is March 2023. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

show abstract

Principled, practical, flexible, fast: a new approach to phylogenetic factor analysis

Cited by 3 publications

References 43 publications

Chameleon biogeographic dispersal is associated with extreme life history strategies

Chameleon biogeographic dispersal is associated with extreme life history strategies

Hamiltonian zigzag accelerates large-scale inference for conditional dependencies between complex biological traits

Data Integration in Bayesian Phylogenetics

Contact Info

Product

Resources

About