Genome-wide association analyses in the model rhizobium Ensifer meliloti

Brendan Epstein, Reda Aboushanab, Abdelaal Shamseldin, Margaret R Taylor, Joseph Guhlin, Liana Burghardt, Matt Nelson, Michael J Sadowsky, Peter L Tiffin

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Genome-wide association studies (GWAS) can identify genetic variants responsible for naturally occurring and quantitative phenotypic variation. Association studies therefore provide a powerful complement to approaches that rely on de novo mutations for characterizing gene function. Although bacteria should be amenable to GWAS, few GWAS have been conducted on bacteria, and the extent to which nonindependence among genomic variants (e.g., linkage disequilibrium [LD]) and the genetic architecture of phenotypic traits will affect GWAS performance is unclear. We apply association analyses to identify candidate genes underlying variation in 20 biochemical, growth, and symbiotic phenotypes among 153 strains of Ensifer meliloti. For 11 traits, we find genotype-phenotype associations that are stronger than expected by chance, with the candidates in relatively small linkage groups, indicating that LD does not preclude resolving association candidates to relatively small genomic regions. The significant candidates show an enrichment for nucleotide polymorphisms (SNPs) over gene presence-absence variation (PAV), and for five traits, candidates are enriched in large linkage groups, a possible signature of epistasis. Many of the variants most strongly associated with symbiosis phenotypes were in genes previously identified as being involved in nitrogen fixation or nodulation. For other traits, apparently strong associations were not stronger than the range of associations detected in permuted data. In sum, our data show that GWAS in bacteria may be a powerful tool for characterizing genetic architecture and identifying genes responsible for phenotypic variation. However, careful evaluation of candidates is necessary to avoid false signals of association.

Original languageEnglish (US)
Article numbere00386-18
JournalmSphere
Volume3
Issue number5
DOIs
StatePublished - Sep 1 2018

Fingerprint

Sinorhizobium meliloti
Genome-Wide Association Study
Genes
Linkage Disequilibrium
Bacteria
Melilotus
Phenotype
Nitrogen Fixation
Symbiosis
Genetic Association Studies
Nucleotides
Mutation
Growth

Keywords

  • BSLMM
  • Bacteria
  • Chip heritability
  • GWAS
  • Genetic architecture
  • Genomics
  • Linkage disequilibrium
  • Medicago
  • Rhizobium
  • Sinorhizobium
  • Symbiosis

Cite this

Genome-wide association analyses in the model rhizobium Ensifer meliloti. / Epstein, Brendan; Aboushanab, Reda; Shamseldin, Abdelaal; Taylor, Margaret R; Guhlin, Joseph; Burghardt, Liana; Nelson, Matt; Sadowsky, Michael J; Tiffin, Peter L.

In: mSphere, Vol. 3, No. 5, e00386-18, 01.09.2018.

Research output: Contribution to journalArticle

@article{e4e15a31723b4fd395777d01f9d8c84e,
title = "Genome-wide association analyses in the model rhizobium Ensifer meliloti",
abstract = "Genome-wide association studies (GWAS) can identify genetic variants responsible for naturally occurring and quantitative phenotypic variation. Association studies therefore provide a powerful complement to approaches that rely on de novo mutations for characterizing gene function. Although bacteria should be amenable to GWAS, few GWAS have been conducted on bacteria, and the extent to which nonindependence among genomic variants (e.g., linkage disequilibrium [LD]) and the genetic architecture of phenotypic traits will affect GWAS performance is unclear. We apply association analyses to identify candidate genes underlying variation in 20 biochemical, growth, and symbiotic phenotypes among 153 strains of Ensifer meliloti. For 11 traits, we find genotype-phenotype associations that are stronger than expected by chance, with the candidates in relatively small linkage groups, indicating that LD does not preclude resolving association candidates to relatively small genomic regions. The significant candidates show an enrichment for nucleotide polymorphisms (SNPs) over gene presence-absence variation (PAV), and for five traits, candidates are enriched in large linkage groups, a possible signature of epistasis. Many of the variants most strongly associated with symbiosis phenotypes were in genes previously identified as being involved in nitrogen fixation or nodulation. For other traits, apparently strong associations were not stronger than the range of associations detected in permuted data. In sum, our data show that GWAS in bacteria may be a powerful tool for characterizing genetic architecture and identifying genes responsible for phenotypic variation. However, careful evaluation of candidates is necessary to avoid false signals of association.",
keywords = "BSLMM, Bacteria, Chip heritability, GWAS, Genetic architecture, Genomics, Linkage disequilibrium, Medicago, Rhizobium, Sinorhizobium, Symbiosis",
author = "Brendan Epstein and Reda Aboushanab and Abdelaal Shamseldin and Taylor, {Margaret R} and Joseph Guhlin and Liana Burghardt and Matt Nelson and Sadowsky, {Michael J} and Tiffin, {Peter L}",
year = "2018",
month = "9",
day = "1",
doi = "10.1128/mSphere.00386-18",
language = "English (US)",
volume = "3",
journal = "mSphere",
issn = "2379-5042",
publisher = "American Society for Microbiology",
number = "5",

}

TY - JOUR

T1 - Genome-wide association analyses in the model rhizobium Ensifer meliloti

AU - Epstein, Brendan

AU - Aboushanab, Reda

AU - Shamseldin, Abdelaal

AU - Taylor, Margaret R

AU - Guhlin, Joseph

AU - Burghardt, Liana

AU - Nelson, Matt

AU - Sadowsky, Michael J

AU - Tiffin, Peter L

PY - 2018/9/1

Y1 - 2018/9/1

N2 - Genome-wide association studies (GWAS) can identify genetic variants responsible for naturally occurring and quantitative phenotypic variation. Association studies therefore provide a powerful complement to approaches that rely on de novo mutations for characterizing gene function. Although bacteria should be amenable to GWAS, few GWAS have been conducted on bacteria, and the extent to which nonindependence among genomic variants (e.g., linkage disequilibrium [LD]) and the genetic architecture of phenotypic traits will affect GWAS performance is unclear. We apply association analyses to identify candidate genes underlying variation in 20 biochemical, growth, and symbiotic phenotypes among 153 strains of Ensifer meliloti. For 11 traits, we find genotype-phenotype associations that are stronger than expected by chance, with the candidates in relatively small linkage groups, indicating that LD does not preclude resolving association candidates to relatively small genomic regions. The significant candidates show an enrichment for nucleotide polymorphisms (SNPs) over gene presence-absence variation (PAV), and for five traits, candidates are enriched in large linkage groups, a possible signature of epistasis. Many of the variants most strongly associated with symbiosis phenotypes were in genes previously identified as being involved in nitrogen fixation or nodulation. For other traits, apparently strong associations were not stronger than the range of associations detected in permuted data. In sum, our data show that GWAS in bacteria may be a powerful tool for characterizing genetic architecture and identifying genes responsible for phenotypic variation. However, careful evaluation of candidates is necessary to avoid false signals of association.

AB - Genome-wide association studies (GWAS) can identify genetic variants responsible for naturally occurring and quantitative phenotypic variation. Association studies therefore provide a powerful complement to approaches that rely on de novo mutations for characterizing gene function. Although bacteria should be amenable to GWAS, few GWAS have been conducted on bacteria, and the extent to which nonindependence among genomic variants (e.g., linkage disequilibrium [LD]) and the genetic architecture of phenotypic traits will affect GWAS performance is unclear. We apply association analyses to identify candidate genes underlying variation in 20 biochemical, growth, and symbiotic phenotypes among 153 strains of Ensifer meliloti. For 11 traits, we find genotype-phenotype associations that are stronger than expected by chance, with the candidates in relatively small linkage groups, indicating that LD does not preclude resolving association candidates to relatively small genomic regions. The significant candidates show an enrichment for nucleotide polymorphisms (SNPs) over gene presence-absence variation (PAV), and for five traits, candidates are enriched in large linkage groups, a possible signature of epistasis. Many of the variants most strongly associated with symbiosis phenotypes were in genes previously identified as being involved in nitrogen fixation or nodulation. For other traits, apparently strong associations were not stronger than the range of associations detected in permuted data. In sum, our data show that GWAS in bacteria may be a powerful tool for characterizing genetic architecture and identifying genes responsible for phenotypic variation. However, careful evaluation of candidates is necessary to avoid false signals of association.

KW - BSLMM

KW - Bacteria

KW - Chip heritability

KW - GWAS

KW - Genetic architecture

KW - Genomics

KW - Linkage disequilibrium

KW - Medicago

KW - Rhizobium

KW - Sinorhizobium

KW - Symbiosis

UR - http://www.scopus.com/inward/record.url?scp=85055608856&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055608856&partnerID=8YFLogxK

U2 - 10.1128/mSphere.00386-18

DO - 10.1128/mSphere.00386-18

M3 - Article

VL - 3

JO - mSphere

JF - mSphere

SN - 2379-5042

IS - 5

M1 - e00386-18

ER -