schliessen

Filtern

 

Bibliotheken

Genealogical Working Distributions for Bayesian Model Testing with Phylogenetic Uncertainty

Marginal likelihood estimates to compare models using Bayes factors frequently accompany Bayesian phylogenetic inference. Approaches to estimate marginal likelihoods have garnered increased attention over the past decade. In particular, the introduction of path sampling (PS) and stepping-stone sampl... Full description

Journal Title: Systematic biology 2016-03-01, Vol.65 (2), p.250-264
Main Author: Baele, Guy
Other Authors: Lemey, Philippe , Suchard, Marc A
Format: Electronic Article Electronic Article
Language: English
Subjects:
Quelle: Alma/SFX Local Collection
Publisher: England: Oxford University Press
ID: ISSN: 1063-5157
Link: https://www.ncbi.nlm.nih.gov/pubmed/26526428
Zum Text:
SendSend as email Add to Book BagAdd to Book Bag
Staff View
recordid: cdi_pubmedcentral_primary_oai_pubmedcentral_nih_gov_5009437
title: Genealogical Working Distributions for Bayesian Model Testing with Phylogenetic Uncertainty
format: Article
creator:
  • Baele, Guy
  • Lemey, Philippe
  • Suchard, Marc A
subjects:
  • Analytical estimating
  • Bayes factor
  • Bayes Theorem
  • Bayesian analysis
  • Bayesian inference
  • Classification - methods
  • coalescent model
  • Computer Simulation
  • Datasets
  • Demography
  • Estimators
  • Evolution, Molecular
  • Evolutionary biology
  • Genealogy
  • HIV 1
  • marginal likelihood
  • MCMC
  • Models, Genetic
  • Overestimates
  • Phylogenetics
  • Phylogeny
  • Population genetics
  • Population size
  • Regular
  • Regular Articles
  • Sampling techniques
  • Topology
  • Working distribution
ispartof: Systematic biology, 2016-03-01, Vol.65 (2), p.250-264
description: Marginal likelihood estimates to compare models using Bayes factors frequently accompany Bayesian phylogenetic inference. Approaches to estimate marginal likelihoods have garnered increased attention over the past decade. In particular, the introduction of path sampling (PS) and stepping-stone sampling (SS) into Bayesian phylogenetics has tremendously improved the accuracy of model selection. These sampling techniques are now used to evaluate complex evolutionary and population genetic models on empirical data sets, but considerable computational demands hamper their widespread adoption. Further, when very diffuse, but proper priors are specified for model parameters, numerical issues complicate the exploration of the priors, a necessary step in marginal likelihood estimation using PS or SS. To avoid such instabilities, generalized SS (GSS) has recently been proposed, introducing the concept of "working distributions'' to facilitate—or shorten—the integration process that underlies marginal likelihood estimation. However, the need to fix the tree topology currently limits GSS in a coalescent-based framework. Here, we extend GSS by relaxing the fixed underlying tree topology assumption. To this purpose, we introduce a "working" distribution on the space of genealogies, which enables estimating marginal likelihoods while accommodating phylogenetic uncertainty. We propose two different "working" distributions that help GSS to outperform PS and SS in terms of accuracy when comparing demographic and evolutionary models applied to synthetic data and real-world examples. Further, we show that the use of very diffuse priors can lead to a considerable overestimation in marginal likelihood when using PS and SS, while still retrieving the correct marginal likelihood using both GSS approaches. The methods used in this article are available in BEAST, a powerful userfriendly software package to perform Bayesian evolutionary analyses.
language: eng
source: Alma/SFX Local Collection
identifier: ISSN: 1063-5157
fulltext: fulltext
issn:
  • 1063-5157
  • 1076-836X
url: Link


@attributes
NO1
SEARCH_ENGINEprimo_central_multiple_fe
SEARCH_ENGINE_TYPEPrimo Central Search Engine
RANK2.6693017
LOCALfalse
PrimoNMBib
record
control
sourceidjstor_pubme
recordidTN_cdi_pubmedcentral_primary_oai_pubmedcentral_nih_gov_5009437
sourceformatXML
sourcesystemPC
jstor_id44028749
oup_id10.1093/sysbio/syv083
sourcerecordid44028749
originalsourceidFETCH-LOGICAL-1550t-41bb14cbfa506c62ccc13e04f9593eb957c566480ed681455f5d15646c992f513
addsrcrecordideNqFkktv1DAURiMEog9YsgRFYtNNwI6f2SCVAgWpPBatQGJhOc7NjIeMPdhOq_x7PGQY0UqIlSPl-Ojzd29RPMHoBUYNeRmn2Fqfj2skyb3iECPBK0n4t_vbb04qhpk4KI5iXCGEMWf4YXFQc1ZzWsvD4vs5ONCDX1ijh_KrDz-sW5RvbEzBtmOy3sWy96F8rSeIVrvyo-9gKC8hpi14Y9Oy_LKcsiB7kjXllTMQkrYuTY-KB70eIjzencfF1bu3l2fvq4vP5x_OTi8qzBhKFcVti6lpe80QN7w2xmACiPYNawi0DROGcU4lgo5LTBnrWYcZp9w0Td0zTI6LV7N3M7Zr6Ay4FPSgNsGudZiU11bd_uPsUi38tWIINZSILPg0C_wGnLYBbt3t8sNUB924UTe9yiWqWmvU1p2gADUY2udwRuJO9sJwaRjNwpNdouB_jrkrtbbRwDBoB36MCos8ByQa0WT0-R105cfgcl9bSmIp-O-E1UyZ4GMM0O8TYqS2e6DmPVDzHmT-2d-V7Ok_g88AuSM0NuntvHNDdvindvcun8v4X4KnM7qKyYc9TCmqpaAN-QVhMtwM
sourcetypeOpen Access Repository
isCDItrue
recordtypearticle
pqid1768187637
display
typearticle
titleGenealogical Working Distributions for Bayesian Model Testing with Phylogenetic Uncertainty
sourceAlma/SFX Local Collection
creatorBaele, Guy ; Lemey, Philippe ; Suchard, Marc A
creatorcontribBaele, Guy ; Lemey, Philippe ; Suchard, Marc A
descriptionMarginal likelihood estimates to compare models using Bayes factors frequently accompany Bayesian phylogenetic inference. Approaches to estimate marginal likelihoods have garnered increased attention over the past decade. In particular, the introduction of path sampling (PS) and stepping-stone sampling (SS) into Bayesian phylogenetics has tremendously improved the accuracy of model selection. These sampling techniques are now used to evaluate complex evolutionary and population genetic models on empirical data sets, but considerable computational demands hamper their widespread adoption. Further, when very diffuse, but proper priors are specified for model parameters, numerical issues complicate the exploration of the priors, a necessary step in marginal likelihood estimation using PS or SS. To avoid such instabilities, generalized SS (GSS) has recently been proposed, introducing the concept of "working distributions'' to facilitate—or shorten—the integration process that underlies marginal likelihood estimation. However, the need to fix the tree topology currently limits GSS in a coalescent-based framework. Here, we extend GSS by relaxing the fixed underlying tree topology assumption. To this purpose, we introduce a "working" distribution on the space of genealogies, which enables estimating marginal likelihoods while accommodating phylogenetic uncertainty. We propose two different "working" distributions that help GSS to outperform PS and SS in terms of accuracy when comparing demographic and evolutionary models applied to synthetic data and real-world examples. Further, we show that the use of very diffuse priors can lead to a considerable overestimation in marginal likelihood when using PS and SS, while still retrieving the correct marginal likelihood using both GSS approaches. The methods used in this article are available in BEAST, a powerful userfriendly software package to perform Bayesian evolutionary analyses.
identifier
0ISSN: 1063-5157
1EISSN: 1076-836X
2DOI: 10.1093/sysbio/syv083
3PMID: 26526428
languageeng
publisherEngland: Oxford University Press
subjectAnalytical estimating ; Bayes factor ; Bayes Theorem ; Bayesian analysis ; Bayesian inference ; Classification - methods ; coalescent model ; Computer Simulation ; Datasets ; Demography ; Estimators ; Evolution, Molecular ; Evolutionary biology ; Genealogy ; HIV 1 ; marginal likelihood ; MCMC ; Models, Genetic ; Overestimates ; Phylogenetics ; Phylogeny ; Population genetics ; Population size ; Regular ; Regular Articles ; Sampling techniques ; Topology ; Working distribution
ispartofSystematic biology, 2016-03-01, Vol.65 (2), p.250-264
rights
0Copyright © 2016 Society of Systematic Biologists
1The Author(s) 2015. Published by Oxford University Press, on behalf of the Society of Systematic Biologists. All rights reserved. For Permissions, please email: journals.permissions@oup.com 2015
2The Author(s) 2015. Published by Oxford University Press, on behalf of the Society of Systematic Biologists. All rights reserved. For Permissions, please email: journals.permissions@oup.com.
lds50peer_reviewed
oafree_for_read
citedbyFETCH-LOGICAL-1550t-41bb14cbfa506c62ccc13e04f9593eb957c566480ed681455f5d15646c992f513
citesFETCH-LOGICAL-1550t-41bb14cbfa506c62ccc13e04f9593eb957c566480ed681455f5d15646c992f513
links
openurl$$Topenurl_article
openurlfulltext$$Topenurlfull_article
thumbnail$$Usyndetics_thumb_exl
backlink$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/26526428$$D View this record in MEDLINE/PubMed
search
creatorcontrib
0Baele, Guy
1Lemey, Philippe
2Suchard, Marc A
title
0Genealogical Working Distributions for Bayesian Model Testing with Phylogenetic Uncertainty
1Systematic biology
addtitleSyst Biol
descriptionMarginal likelihood estimates to compare models using Bayes factors frequently accompany Bayesian phylogenetic inference. Approaches to estimate marginal likelihoods have garnered increased attention over the past decade. In particular, the introduction of path sampling (PS) and stepping-stone sampling (SS) into Bayesian phylogenetics has tremendously improved the accuracy of model selection. These sampling techniques are now used to evaluate complex evolutionary and population genetic models on empirical data sets, but considerable computational demands hamper their widespread adoption. Further, when very diffuse, but proper priors are specified for model parameters, numerical issues complicate the exploration of the priors, a necessary step in marginal likelihood estimation using PS or SS. To avoid such instabilities, generalized SS (GSS) has recently been proposed, introducing the concept of "working distributions'' to facilitate—or shorten—the integration process that underlies marginal likelihood estimation. However, the need to fix the tree topology currently limits GSS in a coalescent-based framework. Here, we extend GSS by relaxing the fixed underlying tree topology assumption. To this purpose, we introduce a "working" distribution on the space of genealogies, which enables estimating marginal likelihoods while accommodating phylogenetic uncertainty. We propose two different "working" distributions that help GSS to outperform PS and SS in terms of accuracy when comparing demographic and evolutionary models applied to synthetic data and real-world examples. Further, we show that the use of very diffuse priors can lead to a considerable overestimation in marginal likelihood when using PS and SS, while still retrieving the correct marginal likelihood using both GSS approaches. The methods used in this article are available in BEAST, a powerful userfriendly software package to perform Bayesian evolutionary analyses.
subject
0Analytical estimating
1Bayes factor
2Bayes Theorem
3Bayesian analysis
4Bayesian inference
5Classification - methods
6coalescent model
7Computer Simulation
8Datasets
9Demography
10Estimators
11Evolution, Molecular
12Evolutionary biology
13Genealogy
14HIV 1
15marginal likelihood
16MCMC
17Models, Genetic
18Overestimates
19Phylogenetics
20Phylogeny
21Population genetics
22Population size
23Regular
24Regular Articles
25Sampling techniques
26Topology
27Working distribution
issn
01063-5157
11076-836X
fulltexttrue
rsrctypearticle
creationdate2016
recordtypearticle
recordideNqFkktv1DAURiMEog9YsgRFYtNNwI6f2SCVAgWpPBatQGJhOc7NjIeMPdhOq_x7PGQY0UqIlSPl-Ojzd29RPMHoBUYNeRmn2Fqfj2skyb3iECPBK0n4t_vbb04qhpk4KI5iXCGEMWf4YXFQc1ZzWsvD4vs5ONCDX1ijh_KrDz-sW5RvbEzBtmOy3sWy96F8rSeIVrvyo-9gKC8hpi14Y9Oy_LKcsiB7kjXllTMQkrYuTY-KB70eIjzencfF1bu3l2fvq4vP5x_OTi8qzBhKFcVti6lpe80QN7w2xmACiPYNawi0DROGcU4lgo5LTBnrWYcZp9w0Td0zTI6LV7N3M7Zr6Ay4FPSgNsGudZiU11bd_uPsUi38tWIINZSILPg0C_wGnLYBbt3t8sNUB924UTe9yiWqWmvU1p2gADUY2udwRuJO9sJwaRjNwpNdouB_jrkrtbbRwDBoB36MCos8ByQa0WT0-R105cfgcl9bSmIp-O-E1UyZ4GMM0O8TYqS2e6DmPVDzHmT-2d-V7Ok_g88AuSM0NuntvHNDdvindvcun8v4X4KnM7qKyYc9TCmqpaAN-QVhMtwM
startdate20160301
enddate20160301
creator
0Baele, Guy
1Lemey, Philippe
2Suchard, Marc A
general
0Oxford University Press
1Society of Systematic Biologists
scope
0CGR
1CUY
2CVF
3ECM
4EIF
5NPM
6AAYXX
7CITATION
8K9.
97X8
10BOBZL
11CLFQK
125PM
sort
creationdate20160301
titleGenealogical Working Distributions for Bayesian Model Testing with Phylogenetic Uncertainty
authorBaele, Guy ; Lemey, Philippe ; Suchard, Marc A
facets
frbrtype5
frbrgroupidcdi_FETCH-LOGICAL-1550t-41bb14cbfa506c62ccc13e04f9593eb957c566480ed681455f5d15646c992f513
rsrctypearticles
prefilterarticles
languageeng
creationdate2016
topic
0Analytical estimating
1Bayes factor
2Bayes Theorem
3Bayesian analysis
4Bayesian inference
5Classification - methods
6coalescent model
7Computer Simulation
8Datasets
9Demography
10Estimators
11Evolution, Molecular
12Evolutionary biology
13Genealogy
14HIV 1
15marginal likelihood
16MCMC
17Models, Genetic
18Overestimates
19Phylogenetics
20Phylogeny
21Population genetics
22Population size
23Regular
24Regular Articles
25Sampling techniques
26Topology
27Working distribution
toplevel
0peer_reviewed
1online_resources
creatorcontrib
0Baele, Guy
1Lemey, Philippe
2Suchard, Marc A
collection
0Medline
1MEDLINE
2MEDLINE (Ovid)
3MEDLINE
4MEDLINE
5PubMed
6CrossRef
7ProQuest Health & Medical Complete (Alumni)
8MEDLINE - Academic
9OpenAIRE (Open Access)
10OpenAIRE
11PubMed Central (Full Participant titles)
jtitleSystematic biology
delivery
delcategoryRemote Search Resource
fulltextfulltext
addata
au
0Baele, Guy
1Lemey, Philippe
2Suchard, Marc A
formatjournal
genrearticle
ristypeJOUR
atitleGenealogical Working Distributions for Bayesian Model Testing with Phylogenetic Uncertainty
jtitleSystematic biology
addtitleSyst Biol
date2016-03-01
risdate2016
volume65
issue2
spage250
epage264
pages250-264
issn1063-5157
eissn1076-836X
notesAssociate Editor: Mark Holder
abstractMarginal likelihood estimates to compare models using Bayes factors frequently accompany Bayesian phylogenetic inference. Approaches to estimate marginal likelihoods have garnered increased attention over the past decade. In particular, the introduction of path sampling (PS) and stepping-stone sampling (SS) into Bayesian phylogenetics has tremendously improved the accuracy of model selection. These sampling techniques are now used to evaluate complex evolutionary and population genetic models on empirical data sets, but considerable computational demands hamper their widespread adoption. Further, when very diffuse, but proper priors are specified for model parameters, numerical issues complicate the exploration of the priors, a necessary step in marginal likelihood estimation using PS or SS. To avoid such instabilities, generalized SS (GSS) has recently been proposed, introducing the concept of "working distributions'' to facilitate—or shorten—the integration process that underlies marginal likelihood estimation. However, the need to fix the tree topology currently limits GSS in a coalescent-based framework. Here, we extend GSS by relaxing the fixed underlying tree topology assumption. To this purpose, we introduce a "working" distribution on the space of genealogies, which enables estimating marginal likelihoods while accommodating phylogenetic uncertainty. We propose two different "working" distributions that help GSS to outperform PS and SS in terms of accuracy when comparing demographic and evolutionary models applied to synthetic data and real-world examples. Further, we show that the use of very diffuse priors can lead to a considerable overestimation in marginal likelihood when using PS and SS, while still retrieving the correct marginal likelihood using both GSS approaches. The methods used in this article are available in BEAST, a powerful userfriendly software package to perform Bayesian evolutionary analyses.
copEngland
pubOxford University Press
pmid26526428
doi10.1093/sysbio/syv083
oafree_for_read