schliessen

Filtern

 

Bibliotheken

Phylogenies and Diversification Rates: Variance Cannot Be Ignored

Abstract A recent pair of articles published in the journal Evolution presented a test for assessing the validity of hierarchical macroevolutionary models. The premise of the test is to compare numerical point estimates of parameters from two levels of analysis; if the estimates differ, the hierarch... Full description

Journal Title: Systematic biology 2019-05-01, Vol.68 (3), p.538-550
Main Author: Rabosky, Daniel L
Format: Electronic Article Electronic Article
Language: English
Subjects:
Publisher: England: Oxford University Press
ID: ISSN: 1063-5157
Link: https://www.ncbi.nlm.nih.gov/pubmed/30481343
Zum Text:
SendSend as email Add to Book BagAdd to Book Bag
Staff View
recordid: cdi_proquest_miscellaneous_2138640704
title: Phylogenies and Diversification Rates: Variance Cannot Be Ignored
format: Article
creator:
  • Rabosky, Daniel L
subjects:
  • Bayes Theorem
  • Biodiversity
  • Classification
  • Computer Simulation
  • Genetic Speciation
  • Models, Biological
  • Models, Statistical
  • Phylogeny
ispartof: Systematic biology, 2019-05-01, Vol.68 (3), p.538-550
description: Abstract A recent pair of articles published in the journal Evolution presented a test for assessing the validity of hierarchical macroevolutionary models. The premise of the test is to compare numerical point estimates of parameters from two levels of analysis; if the estimates differ, the hierarchical model is purportedly flawed. The articles in question apply their proposed test to BAMM, a scientific software program that uses a Bayesian mixture model to estimate rates of evolution from phylogenetic trees. The authors use BAMM to estimate rates from large phylogenies ($n > 60$ tips) and they apply the method separately to subclades within those phylogenies (median size: $n = 3$ tips); they find that point estimates of rates differ between these levels and conclude that the method is flawed, but they do not test whether the observed differences are statistically meaningful. There is no consideration of sampling variation and its impact at any level of their analysis. Here, I show that numerical differences across groups that they report are fully explained by a failure to account for sampling variation in their point estimates. Variance in evolutionary rate estimates—from BAMM and all other methods—is an inverse function of clade size; this variance is extreme for clades with five or fewer tips (e.g., 70% of clades in the focal study). The articles in question rely on negative results that are easily explained by low statistical power to reject their preferred null hypothesis, and this low power is a trivial consequence of high variance in their point estimates. I describe additional mathematical and statistical mistakes that render the proposed testing framework invalid on first principles. Evolutionary rates are no different than any other population parameters we might wish to estimate, and biologists should use the training and tools already at their disposal to avoid erroneous results that follow from the neglect of variance. [Confidence interval; diversification; estimation; null hypothesis; phylogeny; statistics; type I error]
language: eng
source:
identifier: ISSN: 1063-5157
fulltext: no_fulltext
issn:
  • 1063-5157
  • 1076-836X
url: Link


@attributes
NO1
SEARCH_ENGINEprimo_central_multiple_fe
SEARCH_ENGINE_TYPEPrimo Central Search Engine
RANK2.592872
LOCALfalse
PrimoNMBib
record
control
sourceidproquest_opena
recordidTN_cdi_proquest_miscellaneous_2138640704
sourceformatXML
sourcesystemPC
oup_id10.1093/sysbio/syy079
sourcerecordid2138640704
originalsourceidFETCH-LOGICAL-c364t-6ea9f16753117aa37a9e7e4f3620f78790497aa74aa2a2d093053c87923841b50
addsrcrecordideNqFkc1LxDAQxYMofh-9So9eqkmTJq03Xb8RFFHxFqbNRCO7yZp0hf3vrXZdj55meDx-j5lHyB6jh4zW_CjNU-NCP-ZU1Stkk1El84rLl9XvXfK8ZKXaIFspvVPKmCzZOtngVFSMC75JTu7f5uPwit5hysCb7Mx9YkzOuhY6F3z2AB2m4-wZogPfYjYC70OXnWJ2_epDRLND1iyME-4u5jZ5ujh_HF3lt3eX16OT27zlUnS5RKgtk6rkjCkArqBGhcJyWVCrKlVTUfe6EgAFFKY_jZa87fWCV4I1Jd0mNwM3TNGDi6in0U0gznUAp43HThvooHUdalYoUwmoGmsF2ooDZ8pACY2SvDFU9LCDATaN4WOGqdMTl1ocj8FjmCVdMF5JQdWPNR-sbQwpRbTLXEb1dwV6qEAPFfT-_QV61kzQLN2_P__LDrPpP6wvRgqQkw
sourcetypeOpen Access Repository
isCDItrue
recordtypearticle
pqid2138640704
display
typearticle
titlePhylogenies and Diversification Rates: Variance Cannot Be Ignored
creatorRabosky, Daniel L
contributorNear, Thomas
creatorcontribRabosky, Daniel L ; Near, Thomas
descriptionAbstract A recent pair of articles published in the journal Evolution presented a test for assessing the validity of hierarchical macroevolutionary models. The premise of the test is to compare numerical point estimates of parameters from two levels of analysis; if the estimates differ, the hierarchical model is purportedly flawed. The articles in question apply their proposed test to BAMM, a scientific software program that uses a Bayesian mixture model to estimate rates of evolution from phylogenetic trees. The authors use BAMM to estimate rates from large phylogenies ($n > 60$ tips) and they apply the method separately to subclades within those phylogenies (median size: $n = 3$ tips); they find that point estimates of rates differ between these levels and conclude that the method is flawed, but they do not test whether the observed differences are statistically meaningful. There is no consideration of sampling variation and its impact at any level of their analysis. Here, I show that numerical differences across groups that they report are fully explained by a failure to account for sampling variation in their point estimates. Variance in evolutionary rate estimates—from BAMM and all other methods—is an inverse function of clade size; this variance is extreme for clades with five or fewer tips (e.g., 70% of clades in the focal study). The articles in question rely on negative results that are easily explained by low statistical power to reject their preferred null hypothesis, and this low power is a trivial consequence of high variance in their point estimates. I describe additional mathematical and statistical mistakes that render the proposed testing framework invalid on first principles. Evolutionary rates are no different than any other population parameters we might wish to estimate, and biologists should use the training and tools already at their disposal to avoid erroneous results that follow from the neglect of variance. [Confidence interval; diversification; estimation; null hypothesis; phylogeny; statistics; type I error]
identifier
0ISSN: 1063-5157
1EISSN: 1076-836X
2DOI: 10.1093/sysbio/syy079
3PMID: 30481343
languageeng
publisherEngland: Oxford University Press
subjectBayes Theorem ; Biodiversity ; Classification ; Computer Simulation ; Genetic Speciation ; Models, Biological ; Models, Statistical ; Phylogeny
ispartofSystematic biology, 2019-05-01, Vol.68 (3), p.538-550
rights
0The Author(s) 2018. Published by Oxford University Press, on behalf of the Society of Systematic Biologists. All rights reserved. For permissions, please email: journals.permissions@oup.com 2018
1The Author(s) 2018. 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
citesFETCH-LOGICAL-c364t-6ea9f16753117aa37a9e7e4f3620f78790497aa74aa2a2d093053c87923841b50
links
openurl$$Topenurl_article
thumbnail$$Usyndetics_thumb_exl
backlink$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/30481343$$D View this record in MEDLINE/PubMed
search
contributorNear, Thomas
creatorcontribRabosky, Daniel L
title
0Phylogenies and Diversification Rates: Variance Cannot Be Ignored
1Systematic biology
addtitleSyst Biol
descriptionAbstract A recent pair of articles published in the journal Evolution presented a test for assessing the validity of hierarchical macroevolutionary models. The premise of the test is to compare numerical point estimates of parameters from two levels of analysis; if the estimates differ, the hierarchical model is purportedly flawed. The articles in question apply their proposed test to BAMM, a scientific software program that uses a Bayesian mixture model to estimate rates of evolution from phylogenetic trees. The authors use BAMM to estimate rates from large phylogenies ($n > 60$ tips) and they apply the method separately to subclades within those phylogenies (median size: $n = 3$ tips); they find that point estimates of rates differ between these levels and conclude that the method is flawed, but they do not test whether the observed differences are statistically meaningful. There is no consideration of sampling variation and its impact at any level of their analysis. Here, I show that numerical differences across groups that they report are fully explained by a failure to account for sampling variation in their point estimates. Variance in evolutionary rate estimates—from BAMM and all other methods—is an inverse function of clade size; this variance is extreme for clades with five or fewer tips (e.g., 70% of clades in the focal study). The articles in question rely on negative results that are easily explained by low statistical power to reject their preferred null hypothesis, and this low power is a trivial consequence of high variance in their point estimates. I describe additional mathematical and statistical mistakes that render the proposed testing framework invalid on first principles. Evolutionary rates are no different than any other population parameters we might wish to estimate, and biologists should use the training and tools already at their disposal to avoid erroneous results that follow from the neglect of variance. [Confidence interval; diversification; estimation; null hypothesis; phylogeny; statistics; type I error]
subject
0Bayes Theorem
1Biodiversity
2Classification
3Computer Simulation
4Genetic Speciation
5Models, Biological
6Models, Statistical
7Phylogeny
issn
01063-5157
11076-836X
fulltextfalse
rsrctypearticle
creationdate2019
recordtypearticle
recordideNqFkc1LxDAQxYMofh-9So9eqkmTJq03Xb8RFFHxFqbNRCO7yZp0hf3vrXZdj55meDx-j5lHyB6jh4zW_CjNU-NCP-ZU1Stkk1El84rLl9XvXfK8ZKXaIFspvVPKmCzZOtngVFSMC75JTu7f5uPwit5hysCb7Mx9YkzOuhY6F3z2AB2m4-wZogPfYjYC70OXnWJ2_epDRLND1iyME-4u5jZ5ujh_HF3lt3eX16OT27zlUnS5RKgtk6rkjCkArqBGhcJyWVCrKlVTUfe6EgAFFKY_jZa87fWCV4I1Jd0mNwM3TNGDi6in0U0gznUAp43HThvooHUdalYoUwmoGmsF2ooDZ8pACY2SvDFU9LCDATaN4WOGqdMTl1ocj8FjmCVdMF5JQdWPNR-sbQwpRbTLXEb1dwV6qEAPFfT-_QV61kzQLN2_P__LDrPpP6wvRgqQkw
startdate20190501
enddate20190501
creatorRabosky, Daniel L
general
0Oxford University Press
1Oxford University Press (OUP)
scope
0CGR
1CUY
2CVF
3ECM
4EIF
5NPM
6AAYXX
7CITATION
87X8
9CLFQK
sort
creationdate20190501
titlePhylogenies and Diversification Rates: Variance Cannot Be Ignored
authorRabosky, Daniel L
facets
frbrtype5
frbrgroupidcdi_FETCH-LOGICAL-c364t-6ea9f16753117aa37a9e7e4f3620f78790497aa74aa2a2d093053c87923841b50
rsrctypearticles
prefilterarticles
languageeng
creationdate2019
topic
0Bayes Theorem
1Biodiversity
2Classification
3Computer Simulation
4Genetic Speciation
5Models, Biological
6Models, Statistical
7Phylogeny
toplevelpeer_reviewed
creatorcontribRabosky, Daniel L
collection
0Medline
1MEDLINE
2MEDLINE (Ovid)
3MEDLINE
4MEDLINE
5PubMed
6CrossRef
7MEDLINE - Academic
8OpenAIRE
jtitleSystematic biology
delivery
delcategoryRemote Search Resource
fulltextno_fulltext
addata
auRabosky, Daniel L
formatjournal
genrearticle
ristypeJOUR
atitlePhylogenies and Diversification Rates: Variance Cannot Be Ignored
jtitleSystematic biology
addtitleSyst Biol
date2019-05-01
risdate2019
volume68
issue3
spage538
epage550
pages538-550
issn1063-5157
eissn1076-836X
abstractAbstract A recent pair of articles published in the journal Evolution presented a test for assessing the validity of hierarchical macroevolutionary models. The premise of the test is to compare numerical point estimates of parameters from two levels of analysis; if the estimates differ, the hierarchical model is purportedly flawed. The articles in question apply their proposed test to BAMM, a scientific software program that uses a Bayesian mixture model to estimate rates of evolution from phylogenetic trees. The authors use BAMM to estimate rates from large phylogenies ($n > 60$ tips) and they apply the method separately to subclades within those phylogenies (median size: $n = 3$ tips); they find that point estimates of rates differ between these levels and conclude that the method is flawed, but they do not test whether the observed differences are statistically meaningful. There is no consideration of sampling variation and its impact at any level of their analysis. Here, I show that numerical differences across groups that they report are fully explained by a failure to account for sampling variation in their point estimates. Variance in evolutionary rate estimates—from BAMM and all other methods—is an inverse function of clade size; this variance is extreme for clades with five or fewer tips (e.g., 70% of clades in the focal study). The articles in question rely on negative results that are easily explained by low statistical power to reject their preferred null hypothesis, and this low power is a trivial consequence of high variance in their point estimates. I describe additional mathematical and statistical mistakes that render the proposed testing framework invalid on first principles. Evolutionary rates are no different than any other population parameters we might wish to estimate, and biologists should use the training and tools already at their disposal to avoid erroneous results that follow from the neglect of variance. [Confidence interval; diversification; estimation; null hypothesis; phylogeny; statistics; type I error]
copEngland
pubOxford University Press
pmid30481343
doi10.1093/sysbio/syy079