CoChat

Assessing Parameter Identifiability in Phylogenetic Models Using Data Cloning

Open access

The DOI matches its Crossref record; 87 of the 87 references it cites were found in a registry or an index.

Abstract

The success of model-based methods in phylogenetics has motivated much research aimed at generating new, biologically informative models. This new computer-intensive approach to phylogenetics demands validation studies and sound measures of performance. To date there has been little practical guidance available as to when and why the parameters in a particular model can be identified reliably. Here, we illustrate how Data Cloning (DC), a recently developed methodology to compute the maximum likelihood estimates along with their asymptotic variance, can be used to diagnose structural parameter nonidentifiability (NI) and distinguish it from other parameter estimability problems, including when parameters are structurally identifiable, but are not estimable in a given data set (INE), and when parameters are identifiable, and estimable, but only weakly so (WE). The application of the DC theorem uses well-known and widely used Bayesian computational techniques. With the DC approach, practitioners can use Bayesian phylogenetics software to diagnose nonidentifiability. Theoreticians and practitioners alike now have a powerful, yet simple tool to detect nonidentifiability while investigating complex modeling scenarios, where getting closed-form expressions in a probabilistic study is complicated. Furthermore, here we also show how DC can be used as a tool to examine and eliminate the influence of the priors, in particular if the process of prior elicitation is not straightforward. Finally, when applied to phylogenetic inference, DC can be used to study at least two important statistical questions: assessing identifiability of discrete parameters, like the tree topology, and developing efficient sampling methods for computationally expensive posterior densities.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/sysbio/sys055

Cites

87 of 87 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. Posterior distribution for λ1, λ2 and λ1+λ2 in a standard Bayesian analysis (first row) and using DC, with the number of clones r=200 (second row). Note that as the number of clones increases (from no clones in a), b), c) to 200 in d), f), g)), the variance of the posterior distributions for λ1 and λ2 do not converge to 0, yet the variance of the posterior distribution for the sum does converge to 0 (bottom right pannel).
Figure 1. Posterior distribution for λ1, λ2 and λ1+λ2 in a standard Bayesian analysis (first row) and using DC, with the number of clones r=200 (second row). Note that as the number of clones increases (from no clones in a), b), c) to 200 in d), f), g)), the variance of the posterior distributions for λ1 and λ2 do not converge to 0, yet the variance of the posterior distribution for the sum does converge to 0 (bottom right pannel).CC BY-NC · PubMed Central
Figure 2. Empirical and analytical mean and variance of the r-th posterior distribution for the third analytical example with exponential random variables. The first row (panels a) to f)) shows in blue the empirical mean of π(λ1|Y)(r) calculated from MCMC samples of the joint r-th posterior distribution. In red, are the empirical Wald confidence intervals from those MCMC samples, also as a function of the number of clones and as the sample size m of the extra sample that allows identifying separately λ1, decreases to 0. The black crosses are the analytical means of the r-th marginal posterior distribution of λ1. The horizontal black lines denote the location of the analytical ML estimate. The second row [(panels g to l)] shows the changes in the empirical variance from the MCMC samples of the r-th posterior distribution π(r)(λ1|Y) as the number of clones increases, for the same sizes of the second sample. The black crosses denote the corresponding analytical variance of the r-th marginal posterior distribution of λ1
Figure 2. Empirical and analytical mean and variance of the r-th posterior distribution for the third analytical example with exponential random variables. The first row (panels a) to f)) shows in blue the empirical mean of π(λ1|Y)(r) calculated from MCMC samples of the joint r-th posterior distribution. In red, are the empirical Wald confidence intervals from those MCMC samples, also as a function of the number of clones and as the sample size m of the extra sample that allows identifying separately λ1, decreases to 0. The black crosses are the analytical means of the r-th marginal posterior distribution of λ1. The horizontal black lines denote the location of the analytical ML estimate. The second row [(panels g to l)] shows the changes in the empirical variance from the MCMC samples of the r-th posterior distribution π(r)(λ1|Y) as the number of clones increases, for the same sizes of the second sample. The black crosses denote the corresponding analytical variance of the r-th marginal posterior distribution of λ1CC BY-NC · PubMed Central
Figure 3. NMDS representation in Euclidean space of a sample of size 500 trees from the r-th posterior distribution in geodesic space of the tree topologies, as the number of clones increase [(panels a) to e)], for the complete chloroplast data set (86 genes). Each color represents a different topology, and the number of different topologies is specified above each plot. Panel f) depicts the reduction in the relative total variance as the number of clones increases (see text for details).
Figure 3. NMDS representation in Euclidean space of a sample of size 500 trees from the r-th posterior distribution in geodesic space of the tree topologies, as the number of clones increase [(panels a) to e)], for the complete chloroplast data set (86 genes). Each color represents a different topology, and the number of different topologies is specified above each plot. Panel f) depicts the reduction in the relative total variance as the number of clones increases (see text for details).CC BY-NC · PubMed Central
Figure 4. NMDS representation in Euclidean space of a sample of size 500 trees from the r-th posterior distribution in geodesic space of the tree topologies, as the number of clones increase [panels a) to f)], for a single gene (atpB) from the chloroplast data set. Each color represents a different topology, and the number of different topologies is specified above each plot. The insert in panel f) depicts the reduction in the relative total variance as the number of clones increases (see text for details).
Figure 4. NMDS representation in Euclidean space of a sample of size 500 trees from the r-th posterior distribution in geodesic space of the tree topologies, as the number of clones increase [panels a) to f)], for a single gene (atpB) from the chloroplast data set. Each color represents a different topology, and the number of different topologies is specified above each plot. The insert in panel f) depicts the reduction in the relative total variance as the number of clones increases (see text for details).CC BY-NC · PubMed Central
Figure 5. Nonparametric bootstrap distribution of the first eigenvalue of the variance–covariance matrix of the r-th cloned posterior distribution, for r=1, 2, 4, 8, 16, 32 [panels b) to h)], when the resampled replicates are 14, 28, 56,112,225,449, and 898 bp long. Panel a) depicts the rate of decrease of the first eigenvalue of the posterior distribution sample variance for the continuous parameters (in blue) and the trees (in red, see explanation of Figure 4 in the text), and compared with the theoretical rate decrease of 1/r (in black). In panel b), the changes in the first eigenvalue for two other data sets also of size 14 bp are followed with a red and with a blue dotted line. In one of the cases (marked with a blue line), the size of the first eigenvalue decreases sharply as the number of clones increases. In the other case (red line), it does not.
Figure 5. Nonparametric bootstrap distribution of the first eigenvalue of the variance–covariance matrix of the r-th cloned posterior distribution, for r=1, 2, 4, 8, 16, 32 [panels b) to h)], when the resampled replicates are 14, 28, 56,112,225,449, and 898 bp long. Panel a) depicts the rate of decrease of the first eigenvalue of the posterior distribution sample variance for the continuous parameters (in blue) and the trees (in red, see explanation of Figure 4 in the text), and compared with the theoretical rate decrease of 1/r (in black). In panel b), the changes in the first eigenvalue for two other data sets also of size 14 bp are followed with a red and with a blue dotted line. In one of the cases (marked with a blue line), the size of the first eigenvalue decreases sharply as the number of clones increases. In the other case (red line), it does not.CC BY-NC · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 87 references

Research works printed but not linked: 4

  1. Felsenstein, 2004, Inferring phytogenies.
  2. Robert, 1993, Prior feedback: a Bayesian approach to maximum likelihood estimation, Comput. Statist, 8, 279
  3. Robert, 2007, The Bayesian choice: from decision-theoretic foundations to computational implementation.
  4. Ronquist, 2005, MrBayes 3.1 manual.

Cited by

See all 47

Metadata

Authors
Ponciano, J. M., Burleigh, J. G., Braun, E. L., & Taper, M. L.
Affiliations
University of Florida (US)
Corresponding author
Ponciano, José Miguel
Venue
Systematic Biology
Published
2012-07-09
Volume
61
Issue
6
Pages
955-972
ISSN
  • 1063-5157 (ISSN-L)
  • 1076-836X
OpenAlex
W2098284380
Cited by
39
References
86
Citations by year
2025: 2 · 2024: 1 · 2023: 5 · 2022: 3 · 2021: 2 · 2020: 3 · 2019: 7 · 2018: 2 · 2017: 4 · 2016: 1 · 2015: 2 · 2014: 4 · 2013: 3
FWCI
2.75
Type
journal article
Language
en
Open access
Locations
  • Systematic Biology (publishedVersion, cc-by-nc)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
Funders
  • University of Florida
  • College of Liberal Arts and Sciences, University of Florida
Topics
  • Physical Sciences › Mathematics › Statistics and Probability › Markov Chains and Monte Carlo Methods
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Physical Sciences › Mathematics › Statistics and Probability › Statistical Methods and Bayesian Inference
Keywords
parameter identifiability, data cloning, phylogenetic models, estimability, maximum likelihood estimation, asymptotic variance, prior sensitivity, prior elicitation, tree topology, Bayesian phylogenetics, posterior sampling
MeSH terms
Chloroplasts, Classification, Data Interpretation, Statistical, Phylogeny, Plants, Models, Statistical

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.