Genomic Conflict and DNA Sequence Variation

1
Genomic Conflict and DNA Sequence Variation
Marcy K. Uyenoyama Department of Biology Duke
University
2
Overview

• Population genetics
• Historically model-rich
• Present need model-based interpretation of
  observed patterns of genomic variation
• What are hallmarks of each model?
• Self-incompatibility systems in plants
• Recognizing genomic conflict due to sexual
  antagonism

3
Canonical models

• Neutral evolution
• Pure neutrality distribution of offspring
  number is independent of any trait in parent
• Demographic history deme founding, gene flow
• Purifying selection maintain functioning state
  against random deleterious mutations
• Selection
• Balancing selection maintenance of different
  forms
• Selective sweeps substitution of most fit for
  less fit

4
Hallmarks of evolution

• How do we know it when we see it?
• Patterns evident in genome variation
• Model selection
• Choosing among a small number of canonical models
  for any particular system

5
A random sample of genes
Observed
Sample
Ancestral sequence
6
Allele and mutation spectra
Site frequency spectrum
Number of mutations
Multiplicity
a a1 6, a3 1, a5 1, a6 1, for ai the
number of alleles with multiplicity i
7
The neutral coalescent
Sample root from stationary distribution of
P,mutation transition matrix and bifurcate

• After an interval choose a lineage at random
• Replace it by two identical copies with
  probability
• Mutate it according to P with probability

8
Evolutionary rates

• Events on level k
• Bifurcation at rate Mutation at rate
• Population parameters ratios of rates
• Next event is a bifurcation/coalescence with
  probability

9
Allele and mutation spectra
Site frequency spectrum
Number of mutations
Multiplicity
a a1 6, a3 1, a5 1, a6 1, for ai the
number of alleles with multiplicity i
10
Infinite-alleles model

• Mutation
• Novel allelic types formed at rate u per gene per
  generation
• Reproduction
• Frequency of allele i in the parental population
  pi
• Multinomial sampling of N genes to form the
  offspring
• To find probability of the sample of n genes
  (n1, n2, , nk) or (a1, a2, , an)
• for k the number of distinct haplotypes
  (alleles) ni the number of replicates of allele
  i ai the number of alleles with i replicates

11
Ewens sampling formula
a (a1, a2, , an), for ai the number of alleles
represented by i replicates in a sample of size
n ? 2Nu, for N the effective number of genes
and u the per-locus, per-generation rate of
mutation
Ewens (1972, Theoretical Population Biology)
12
Allele and mutation spectra
Site frequency spectrum
Number of mutations
Multiplicity
a a1 6, a3 1, a5 1, a6 1, for ai the
number of alleles with multiplicity i
13
Population genomics
About 750 accessions isolated from natural
populations worldwide Summary statistics for
sample of 19 entire genomes
http//www.arabidopsis.org
14
Arabidopsis SNP spectra
2
Minor allele counts
3
5
6
7
8
4
Site frequency spectra differ among functional
classes
Kim et al. (2008 Nature Genetics. 39 1151)
15
ESF conditioned on two alleles

• Biallelic sample of size m
• Multiplicities i and (m i )

independent of ?!
16
Ewens sampling formula
a (a1, a2, , an), for ai the number of alleles
represented by i replicates in a sample of size
n ? 2Nu, for N the effective number of genes
and u the per-locus, per-generation rate of
mutation
Ewens (1972, Theoretical Population Biology)
17
Actual site frequency spectra
Excess of rare and common types, deficiency of
intermediate types Data from NIEHS Environmental
Genome Project Direct resequencing of loci
considered environmentally-sensitive Global
representation of ethnicities
Hernandez, Williamson, and Bustamante (2007)
18
Spectrum shape
Signature of expansion? Expansions maintain more
rare mutations Signature of selective
sweep? Neutral variants experience selection asa
population bottleneck
Black constant population size Grey recent
expansion from small population size
19
Arabidopsis SNP spectra
2
Minor allele counts
3
5
6
7
8
4
Site frequency spectra differ among functional
classes
Kim et al. (2008 Nature Genetics. 39 1151)
20
Modelling a SNP data set
Nordborg (2001 Handbook of Statistical Genetics)

• Single segregating mutation in the sample
  genealogy
• Conditional on exactly one segregating site,
  determine the distribution of the size (number of
  descendants) of the branch on which the mutation
  occurs
• Exactly two alleles in the sample
• Conditional on two haplotypes, bearing any number
  of segregating sites, determine the distribution
  of numbers of the two alleles

21
Conditioning

• Two alleles
• One segregating site

22
Multiplicity conditioned on a SNP

• Single segregating site in a sample of size m
• Multiplicity i

dependent on ? !
Ganapathy and Uyenoyama (2009 Theoretical
Population Biology)
23
Arabidopsis SNP spectra
2
Minor allele counts
3
5
6
7
8
4
Site frequency spectra differ among functional
classes
Kim et al. (2008 Nature Genetics. 39 1151)
24
Overview

• Population genetics
• Historically model-rich
• Present need model-based interpretation of
  observed patterns of genomic variation
• What are hallmarks of each model?
• Self-incompatibility systems in plants
• Recognizing genomic conflict due to sexual
  antagonism

25
Genomic conflict

• Phenotypes
• Multiple genes generally influence a given
  phenotype
• Conflict
• Target trait value differs among genes that
  control phenotype
• Sexual antagonism
• Male and female function collaborate in
  reproduction
• Genes influencing each function may come into
  conflict

26
Conflict and genomic variation

• Mating type regions as a battleground
• S-locus controls self-incompatibility in
  flowering plants
• How does sexual antagonism affect the pattern of
  molecular-level variation within the S-locus?
• What are hallmarks of conflict?
• Develop a basis for inference
• Model-based approach to the analysis of genetic
  variation

27

• Flower development
• Basic perfect flower includes both male and
  female components
• Fertilization
• Pollen grains deposited on stigma germinate and
  pollen tubes grow down style to the ovary

Mariana Ruiz http//commons.wikimedia.org/wiki/Fil
eMature_flower_diagram.svg
28
Mariana Ruiz http//commons.wikimedia.org/wiki/Fil
eMature_flower_diagram.svg

• Gametophytic SI (GSI)
• Specificity expressed by individual pollen grain
  or tube determined by own S-allele
• Pollen rejection
• Growth of pollen tube arrested in style

Norbert Holstein http//commons.wikimedia.org/wiki
/FileGametophytic_self-incompatibility.png
29
Mariana Ruiz http//commons.wikimedia.org/wiki/Fil
eMature_flower_diagram.svg

• Sporophytic SI (SSI)
• Specificity expressed by individual pollen grain
  or tube determined by the S-locus genotype of its
  parent
• Pollen rejection
• Germination of pollen grain may be arrested at
  stigma surface

Norbert Holstein http//commons.wikimedia.org/wiki
/FileSporophytic_self-incompatibility.png
30
Mariana Ruiz http//commons.wikimedia.org/wiki/Fil
eMature_flower_diagram.svg
Norbert Holstein http//commons.wikimedia.org/wiki
/FileGametophytic_self-incompatibility.png
Pistil (A) component rejection ofrecognized
specificities Pollen (B) component declaration
ofspecificity
Norbert Holstein http//commons.wikimedia.org/wiki
/FileSporophytic_self-incompatibility.png
31
Mating type regions
Uyenoyama (2005)
32
Human Y chromosome
Skaletsky et al. (2003 Nature 423 825)

• Non-recombining male-specific Y (MSY)
• Euchromatic region 23 MB
• Differences between two random Ys every 3 4 KB
• Mammalian sex determinant SRY
• Y-linked regulator of transcription of many
  male-specific Y-linked genes

33
Mating type regions

• Linkage between pistil (A) and pollen
  (B)components is essential to SI function
• Pollen declaration of specificity
• Pistil rejection of recognized specificities

Uyenoyama (2005)
34
Brassica S-locus
Natural populations often contain 30 50
S-alleles
Nasrallah (2000 Curr. Opin. Plant Biol.)
35
Ubiquitin tags proteins for degradation

• Style S-RNase disrupts pollen tube growth
• Upon entering a pollen tube, S-RNases initially
  sequestered in a vacuole
• In incompatible crosses, vacuole breaks down,
  releasing S-RNases into cytoplasm of pollen tube
• Pollen SLF (S-locus F-box)
• Mediator of ubiquitinylation (attachment of
  ubiquitin)
• Disables all S-RNases except those of the same
  specificity

Vierstra (2009, Nature Reviews Molecular Cell
Biology)
36
Sexual antagonism

• Pistil why reject fertilization?
• Screening of potential mates may improve
  offspring quality
• Cost under incomplete reproductive compensation
  ovules may go unfertilized
• Pollen why provoke rejection?
• Self-rejection may improve quality of own ovules
• Rejection by other plants reduces siring success
• Hide behind another S-specificity in sporophytic
  SI?
• Decline to declare S-specificity altogether?

37
GSI model

• Basic discrete time recursion
• Symmetries in genotype and allele frequencies
• Model change in frequency of focal allele i,
  assuming all other alleles in equal frequency

Wright (1937, Genetics)
38
Diffusion approximation

• Change in allele frequency
• Diffusion equation coefficients
• holds for large population size (N) and u (rate
  of mutation to new S-alleles) of order 1/N

Wright (1937, Genetics)
39
Wrights diffusion model

• Diffusion with jumps
• Turnover rate

40
Expansion of time scale under balancing selection

• High rate of invasion of rare alleles
• Promotes invasion of new and retention of rare
  types
• Maintains high numbers of alleles
• Genealogical relationships
• Tree shape similar under symmetric balancing
  selection and neutrality
• Greatly expanded time scale

Takahata (1993, Mechanisms of Molecular Evolution)
41
S-allele turnover

• Quasi-equilibrium of S-alleles
• Invasion of new, rare S-alleles balanced by
  extinction of common S-alleles
• Expansion of time scale
• Rate of divergence among S-allele classes similar
  to rate among neutral lineages, but in a
  population of size fN

42
Gametophytic SI models

• Basic discrete time recursion
• Diffusion approximation
• Parameters
• Effective population size (N)
• Rate of mutation to new S-specificities (u)

43
Simulation results

• Stationary distribution of allele frequency
• Most time spent close to deterministic
  equilibrium (1/n) or in boundary layer close to
  extinction
• Number of S-alleles
• Analytical expectation for number of common
  S-alleles

Vallejo-Marín and Uyenoyama (2008)
44
Mariana Ruiz http//commons.wikimedia.org/wiki/Fil
eMature_flower_diagram.svg
Norbert Holstein http//commons.wikimedia.org/wiki
/FileGametophytic_self-incompatibility.png
Pistil (A) component rejection ofrecognized
specificities Pollen (B) component declaration
ofspecificity
Norbert Holstein http//commons.wikimedia.org/wiki
/FileSporophytic_self-incompatibility.png
45
Pollen specificity in GSI

• Each pollen expresses its own specificity
• Rarer specificities are incompatible with fewer
  plants
• Incompatible matings
• For n S-alleles in equal frequencies, a pollen
  type is incompatible with a proportion 2/n of all
  plants

Norbert Holstein http//commons.wikimedia.org/wiki
/FileGametophytic_self-incompatibility.png
46
Sexual antagonism

• Pistil why reject fertilization?
• Screening of potential mates may improve
  offspring quality
• Cost under incomplete reproductive compensation
  ovules may go unfertilized
• Pollen why provoke rejection?
• Self-rejection may improve quality of own ovules
• Rejection by other plants reduces siring success
• Hide behind another S-specificity in sporophytic
  SI?
• Decline to declare S-specificity altogether?

47
Fate of style-part mutant
Full SC
Polymorphism
Full SI
48
Fate of pollen-part mutant
Full SC
Relative viability of inbred offspring (s )
Disruption
Polymorphism
Full SI
Self-pollen fraction (s)
Uyenoyama, Zhang, and Newbigin (2001)
49
An
Bn
Sn
Sb
Sa
Sn1
Uyenoyama, Zhang, and Newbigin (2001)
50
An
Bn
TURN OFF Partial breakdown of SIby pollen
disablement
Sn
Evolutionarily unlikely
Sb
Sa
TURN ON Restoration of SIby stylar recognition
Evolutionarily unlikely
Sn1
Uyenoyama, Zhang, and Newbigin (2001)
51
Joint genealogies
Solanaceae and Plantaginaceae
Rosaceae

• Unlike S-RNase genes, SLF genes show
• Low divergence between allelic types
• No trans-specific sharing of lineages

Newbigin, Paape, and Kohn (2008)
52
Cycles of loss/restoration of SI?

• Family-specific genealogies
• Rosaceae do highly-diverged, ancient SFB
  lineages reflect continuous operation or
  restoration of same F-box genes?
• Solanaceae, Plantaginaceae Recruitment of new
  F-box genes?
• Turnover of pollen-specificity loci
• Expression and recognition of a paralogue of the
  former pollen specificity gene?
• Can homologues be distinguished from paralogues
  with new function?

53
Brassica S-locus
Natural populations often contain 30 50
S-alleles
Nasrallah (2000 Curr. Opin. Plant Biol.)
54
An appeal for inference methods

• Sexual antagonism in mating type regions
• Neutral variation in linked regions
• Rates of substitution at determinants of mating
  type
• Inference
• Goal use the pattern of variation in population
  samples of genomic regions as a basis for
  inference about the evolutionary process
• Detection
• genomic conflict and other forms of selection
• mating systems and population structure

55
Pollen specificity in SSI

• Codominance
• Both specificities expressed
• Almost twice as many incompatible styles under
  SSI than GSI for same number of S-alleles
• Complete dominance
• One specificity expressed

Norbert Holstein http//commons.wikimedia.org/wiki
/FileSporophytic_self-incompatibility.png
56
SRK genealogies

• Sporophytic SI
• Diploid genotype of pollen parent determines
  S-specificity of each pollen grain
• Class I is dominant over Class II, with
  codominance within class
• Class II pollen-recessive
• Lower number of segregating alleles, each with
  relatively higher frequency in population
• Greater genealogical relationship within class?

Edh, Widén and Ceplitis (2009)
57
Is class II younger than class I?

• MRCA ages
• Class I 25.5 8.1 MY
• Class II 3.1 0.9 MY
• I/II 41.4 12.7 MY
• Origin of SLG/SRK system
• 42.1 9.0 MY

Uyenoyama (1995)