Umeå University's logo

umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Inference of the distribution of fitness effects of mutations is affected by single nucleotide polymorphism filtering methods, sample size and population structure
Umeå University, Faculty of Science and Technology, Department of Ecology and Environmental Sciences.ORCID iD: 0000-0002-7657-6933
Umeå University, Faculty of Science and Technology, Department of Ecology and Environmental Sciences.ORCID iD: 0000-0001-9437-3198
Department of Computational Biology, Cornell University, NY, Ithaca, United States.
Umeå University, Faculty of Science and Technology, Department of Mathematics and Mathematical Statistics. Advancing Systems Analysis Program, International Institute for Applied Systems Analysis, Laxenburg, Austria; Complexity Science and Evolution Unit, Okinawa Institute of Science and Technology Graduate University, Kunigami, Japan.ORCID iD: 0000-0001-9862-816x
Show others and affiliations
2023 (English)In: Molecular Ecology Resources, ISSN 1755-098X, E-ISSN 1755-0998, Vol. 23, no 7, p. 1589-1603Article in journal (Refereed) Published
Abstract [en]

The distribution of fitness effects (DFE) of new mutations has been of interest to evolutionary biologists since the concept of mutations arose. Modern population genomic data enable us to quantify the DFE empirically, but few studies have examined how data processing, sample size and cryptic population structure might affect the accuracy of DFE inference. We used simulated and empirical data (from Arabidopsis lyrata) to show the effects of missing data filtering, sample size, number of single nucleotide polymorphisms (SNPs) and population structure on the accuracy and variance of DFE estimates. Our analyses focus on three filtering methods—downsampling, imputation and subsampling—with sample sizes of 4–100 individuals. We show that (1) the choice of missing-data treatment directly affects the estimated DFE, with downsampling performing better than imputation and subsampling; (2) the estimated DFE is less reliable in small samples (<8 individuals), and becomes unpredictable with too few SNPs (<5000, the sum of 0- and 4-fold SNPs); and (3) population structure may skew the inferred DFE towards more strongly deleterious mutations. We suggest that future studies should consider downsampling for small data sets, and use samples larger than 4 (ideally larger than 8) individuals, with more than 5000 SNPs in order to improve the robustness of DFE inference and enable comparative analyses.

Place, publisher, year, edition, pages
John Wiley & Sons, 2023. Vol. 23, no 7, p. 1589-1603
Keywords [en]
DFE, missing-data treatment, population structure, sample size, SLiM simulation
National Category
Evolutionary Biology
Identifiers
URN: urn:nbn:se:umu:diva-211803DOI: 10.1111/1755-0998.13825ISI: 001015493200001PubMedID: 37340611Scopus ID: 2-s2.0-85162975600OAI: oai:DiVA.org:umu-211803DiVA, id: diva2:1781928
Funder
Swedish Research CouncilSwedish National Infrastructure for Computing (SNIC)Available from: 2023-07-11 Created: 2023-07-11 Last updated: 2024-07-02Bibliographically approved
In thesis
1. Finding fitness: empirical and theoretical explorations of inferring fitness effects from population-level SNP data
Open this publication in new window or tab >>Finding fitness: empirical and theoretical explorations of inferring fitness effects from population-level SNP data
2024 (English)Doctoral thesis, comprehensive summary (Other academic)
Alternative title[sv]
Jakten på fitness : empiriska och teoretiska undersökningar av hur mutationers fitness-effekter beräknas från SNP-data på populationsnivå
Abstract [en]

The distribution of fitness effects (DFE) describes the likelihood that a new mutation has a specific effect on the fitness of an individual in a given population. The shape of the DFE is a result of several factors such as population size, mating system and selective environment, and can in turn influence the evolutionary potential of a species. The DFE has long been a field of intense research, but particularly since molecular methods enabled us to study of genetic variation in organisms empirically. This research has led to the development of several statistical methods that use population-level frequencies of single nucleotide polymorphisms (SNPs) to infer the DFE. However, these methods rely on assumptions about the data and the organism itself, which could potentially affect the accuracy of the inferences. In this thesis, I describe how two major factors – data quality and inbreeding – can affect the accuracy of DFE inferences. I also show how and when to (and when not to) use DFE inference methods based on SNP frequencies.

All genomic datasets contain inaccuracies and some level of uncertainty. The data sets are therefore often treated to remove the gaps or less reliable information, such as genotypes with low coverage. Some data sets need heavy filtering, which could reduce the amount of data available for analysis. We show that the choice of filter method affects the size of the final data set and the accuracy of the estimated DFE.

Many DFE estimation software assumes random mating within the study population. Unfortunately, this assumption induces some error when trying to estimate the DFE in inbred or selfing species. Some have assumed that this is a result of high rates of homozygosity in the data, and should only be a problem in populations with very high rates of selfing (>99%). We show that accuracy of the estimated DFE decreases already at relatively low rates of selfing (70%) and that removing homozygosity does not improve the accuracy, implying that another mechanism could be causing the error.

Abstract [sv]

Nya mutationer kan ha olika effekt på fitness hos en individ; en mutation kan vara negativ, neutral eller positiv för överlevnad och/eller reproduktion. Sannolikheten att en mutation har en specifik påverkan på fitness kan beskrivas av fördelningen av fitnesseffekter, i vetenskaplig litteratur kallad "the distribution of fitness effects” eller DFE. Formen på DFEn hos en art eller population påverkas av faktorer såsom effektiv populationsstorlek, parningssystem och miljö, och kan i sin tur påverka artens/populationens evolutionspotential. Hur DFEn ser ut hos olika arter har länge varit ett aktivt forskningsfält, och fick ytterligare ett uppsving efter att molekylära metoder gjorde det möjligt att studera genetisk variation empiriskt. Denna utveckling ledde även till att en uppsjö av statistiska metoder utvecklades för att uppskatta DFEn från allelfrekvenser hos punktmutationer i en population. Dessa modeller bygger emellertid på ett antal antaganden om populationen och datan, där osanna antaganden kan orsaka feluppskattningar av DFEn. I denna avhandling undersöker jag hur två faktorer – datakvalitet och inavel – kan påverka hur väl dessa metoder uppskattar den korrekta DFEn. Jag beskriver även hur och när man bör (och inte bör) använda allelefrekvensbaserade metoder för att uppskatta DFE hos en art.

Alla genomiska dataset innehåller en viss grad av osäkerhet och kan sakna information för vissa individer och/eller platser i genomet. Denna sortens data brukar därför förbehandlas för att exkludera obefintliga, osäkra och eventuellt felaktiga data. Vissa behandlingsmetoder kan exkludera stora delar data, beroende på hur mönstret av osäker data ser ut. Jag visar att mängden och magnituden hos feluppskattningar av DFEn beror på både valet av filtreringsmetod och storleken på det slutgiltiga datasetet.

  Många metoder för DFE-uppskattning utgår ifrån teoretiska modeller som antar slumpmässig parning inom studiepopulationen. Detta antagande kan dock introducera feluppskattningar när de används på inavlade och/eller självbefruktande arter. Vissa tidigare studier har påstått att denna effekt beror på inavlade arters höga homozygositet, och bara borde vara ett problem vid stark inavel eller nästan uteslutande (>99%) självbefruktning. Jag beskriver här att feluppskattningar av DFEn blir vanligare redan vid lägre förekomst av självbefruktning (70-80%), samt att homozygositet i sig inte verkar vara den ledande orsaken till feluppskattningar av DFEn. Detta tyder därmed på att någon annan mekanism orsakar de större felmarginaler vi ser vid DFE-analys av inavlade arter.

Place, publisher, year, edition, pages
Umeå University, 2024. p. 45
Keywords
distribution of fitness effects, site frequency spectra, missing data filtering, downsampling, imputation, sample size, population structure, inbreeding, selfing, homozygosity, Arabidopsis, Pinus, DFE-alpha, SLiM, population genomics, adaptation
National Category
Evolutionary Biology
Research subject
biology; evolutionary genetics; Population Biology; Molecular Biology; Genetics
Identifiers
urn:nbn:se:umu:diva-219262 (URN)978-91-8070-269-0 (ISBN)978-91-8070-268-3 (ISBN)
Public defence
2024-02-02, MA121, MIT-huset, Campustorget 5, Umeå, 09:00 (English)
Opponent
Supervisors
Available from: 2024-01-12 Created: 2024-01-10 Last updated: 2024-01-10Bibliographically approved

Open Access in DiVA

fulltext(915 kB)98 downloads
File information
File name FULLTEXT02.pdfFile size 915 kBChecksum SHA-512
c0de420c851ecc81d506737b78e64043f72a3113b68559ba158e43972bb197054e1866c86a5c81525c6a65768767e09580e2b7d00510e850379367217b429d41
Type fulltextMimetype application/pdf

Other links

Publisher's full textPubMedScopus

Authority records

Andersson, BeaZhao, WeiBrännström, ÅkeWang, Xiao-Ru

Search in DiVA

By author/editor
Andersson, BeaZhao, WeiBrännström, ÅkeWang, Xiao-Ru
By organisation
Department of Ecology and Environmental SciencesDepartment of Mathematics and Mathematical Statistics
In the same journal
Molecular Ecology Resources
Evolutionary Biology

Search outside of DiVA

GoogleGoogle Scholar
Total: 127 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

doi
pubmed
urn-nbn

Altmetric score

doi
pubmed
urn-nbn
Total: 459 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf