Transcript labs.bio.unc.edu

Timeline of plant domestication
centuries b.p.
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
Old World
New World
emmer, rice
einkorn
almond
apple, lentil, bean
bean
walnut
spelt. date, broccoli
maize, cotton
cane, chickpea, lettuce, olive, cucumber
chili, squash, potato, avocado
grape, citrus, watermelon
barley, pea, carrot, onion, garlic, fig, tea
peanut
celery, cabbage, artichoke
tomato
beet, banana
eggplant, spinach, coffee
papaya
pecan, cashew, pineapple
The domestication syndrome
 Harvestable parts
 number
 size
 Life history
 determinate flowering
 loss of perenniality
 loss of outcrossing
 Dispersal
 loss of shattering
 loss of seed dormancy
Domestication: a convenient model
to study the genetics adaptation
Strong, long-term artificial selection
Phenotypes are well-characterized
Potential for genetic dissection
maps and markers
sequences
mutant stocks
transformation technologies, etc.
The genetic architecture of domestication:
conventional wisdom
 A few major genes with
recessive gene action
 “Sport” model for
domestication alleles
 Loss of function mutants
 Too deleterious to be
present in wild
populations
 De novo mutations that
arose in cultivated, predomesticated populations
The genetic architecture of domestication:
recent studies
Quantitative trait locus (QTL) mapping
 Cross domesticated genotype with a wild relative
 Follow inheritance of mapped markers in
segregating progeny (F2, backcross, etc.)
 Measure domestication-related phenotypes in the
progeny
 Identify genomic regions cosegregating with a
phenotype → linked to domestication QTL
 Estimate location, effect size, dominance, etc.
M
M
***
Q
M
m
m
m
***
q
m
m
M
M
Q
M
M
Q
M
M
M
M
Q
m
m
m
m
q
M
M
m
q
m
m
m
q
m
m
M
m
X
X
X
X
X
X
m
M
QTL mapping: caveats
Locus number is underestimated
 Only detect loci of largest effect
 Conflate linked genes
Effect size is biased upward in small samples
Dominance
 Cannot be detected in all crosses
 May reflect combined effect of linked loci
Low power to detect interactions
Coarse resolution
Unexpected patterns in domestication QTL:
clustering in sunflower
Unexpected patterns in domestication QTL:
homologous loci in related species
Tomato fruit weight QTL
 highly polygenic trait
 the largest ones map to
homologous genomic regions
in eggplant and pepper
Homologous fruit weight QTL in tomato (Grandillo 1999), pepper (Ben
Chaim et al. 2001) and eggplant (Doganlar et al. 2002). %C is percent
difference of homozygous introgression relative to control. %PVE is
percent phenotypic variance explained. Entries marked by a period (.)
indicate no significant QTL at the syntenic location.
Tomato
Pepper
Eggplant
Locus
%C
Locus
%PVE
Locus
%PVE
fw2.2
20
fw2.1
10
fw2.1
23
fw3.1
13
fw3.2
15
.
.
fw9.2
9
.
.
fw9.1
33
fw11.1
10
.
.
fw11.1
19
Unexpected patterns in domestication QTL:
ancient
Maize tb1
 ‘Domestication’ allele
found in teosinte
Tomato fw2.2 also
diverged prior to
domestication
Clark et al. (2003) PNAS 101: 700
Domestication QTL studies
genus
Capsicum
Citrullus
Gossypium
common name
pepper
watermelon
cotton
Helianthus
Lactuca
Lycopersicon
Oryza
sunflower
lettuce
tomato
rice
Panicum
Phaseolus
Saccharum
Solanum
Sorghum
Triticum
Vigna
Zea
Zizania
pearl millet
common bean
sugarcane
eggplant
sorghum
wheat
cowpea
maize
wildrice
references
Rao et al. 2003
Hashizume et al. 1993
Jiang et al. 1998,
Wright et al. 1999,
Mei et al. 2003
Burke et al. 2002
Johnson et al. 2000
Grandillo & Tanksley 1996
Xiong et al. 1999,
Cai & Morishima, 2000,
Bres-Patry et al. 2001
Poncet et al. 2000, 2002
Koinange et al. 1996
Ming et al. 2001, 2002
Doganlar et al. 2002
Lin et. al. 1995
Peng et al. 2003
Ewa Ubi et al. 2000
Doebley & Stec 1991
Kennard et al. 2002
Questions
 What is the genetic architecture of a typical
domestication trait?
 Major gene
 Polygenic
Origin of domestication alleles
 New mutations or pre-existing alleles?
Clustering of QTL
 Why does it occur?
 Is there a potential relationship to QTL homology?
Questions
 What is the genetic architecture of a typical
domestication trait?
 Major gene
 Polygenic
Origin of domestication alleles
 New mutations or pre-existing alleles?
Clustering of QTL
 Why does it occur?
 Is there a potential relationship to QTL homology?
Experimental design variables
Type of experimental cross
 F2
 Backcross
 Recombinant inbred (no heterozygotes)
 Doubled haploid (no heterozygotes)
Number of individuals
Density (and type) of markers
We reduce all these variables to one
measure: statistical power
Estimating statistical power
Power
 Loosely, the probability of detecting a QTL when it
is present (ranges from 0 to 1)
 Calculated by simulation
Assumptions
 Single codominant QTL
 Constant small additive effect
 Constant environmental variance
Power for a “middle-of-the-road” QTL
 Allows us to compare different experiments on a
common yardstick
Power and number of QTL
20
# QTL detected per trait
18
16
14
12
10
8
6
4
2
0
0
0.2
0.4
0.6
Power
0.8
1
Power and effect size
100
90
80
70
60
50
40
30
20
10
0
0
0.1
0.2
0.3
0.4
0.5
Power
0.6
0.7
0.8
0.9
1
Beavis effect:
QTL effect size
overestimating the effect size of detected QTL
significance threshold
Large
Small
Population size
Results from low power studies
can be misleading
 Numbers of QTL
 Five or more QTL per trait are only detected when power >
0.8
 When power > 0.8, about half the traits have > 5 QTL
 Effect sizes
 QTL of less than 20% PVE are rarely seen unless power >
0.75-0.8
 Most QTL contribute <20% PVE when power is >0.8
 This may be due to both undetected QTL and biased
estimates of effect size
Questions
 What is the genetic architecture of a typical
domestication trait?
 Major gene
 Polygenic
Origin of domestication alleles
 New mutations or pre-existing alleles?
Clustering of QTL
 Why does it occur?
 Is there a potential relationship to QTL homology?
Two models for the origin of
domestication alleles
 ‘Sports’, or new mutations (Lester 1989, Ladizinsky 1998)
 Selected from standing variation (e.g. maize tb1, tomato fw2,2).
 Contrasting the dominance of QTL between selfers and
outcrossers allows us to distinguish these models (Orr and
Betancourt 2001).
 If adaptation uses
 new mutations: selfers will fix more recessive alleles than
outcrossers (consistent with the standard “sport model”)
 standing variation: the probability of fixation of an allele is
independent of dominance (assuming s- balance) and relatively
insensitive to mating system
 We compared the 8 QTL studies where dominance can be
estimated
Gene action
Genotype
A2A2
Genotypic
value
-a
(Underdominant)
-1.25
-1.00
0
Recessive
-0.75
-0.25
Additive
0
A1A2
A1A1
d
+a
Dominant
0.25
Gene action of A1 = d/a
0.75
(Overdominant)
1.00
1.25
Gene action of domestication alleles
Recessive Codominant Dominant Avg. d/a
15
7
22
0.03
16
11
3
0.10
6
1
2
1.40
24
7
16
0.78
67 (46%)
26 (18%)
54 (37%)
0.57
Rice
Tomato
Sorghum
Eggplant
All selfers
Maize
Sunflower
Pearl millet
Wild rice
All outcrossers
Total
17
29
31
16
93 (40%)
15
17
14
13
59 (26%)
11
32
22
13
78 (34%)
160
85
132
Two-tailed t-test of d/a: p<0.31
0.11
0.13
0.25
0.43
0.015
Origin of domestication alleles
The domesticated allele is recessive or
dominant with roughly equal frequency
Dominance does not differ appreciably
between selfers and outcrossers
Results are more compatible with the
predictions of the ‘standing variation’ model
The spectrum of dominance resembles that
seen in surveys of insecticide resistance
alleles (Bourget and Raymond 1998)
Questions
 What is the genetic architecture of a typical
domestication trait?
 Major gene
 Polygenic
Origin of domestication alleles
 New mutations or pre-existing alleles?
Clustering of QTL
 Why does it occur?
 Is there a potential relationship to QTL homology?
QTL clustering:
Three potential explanations
Introgression model
An artifact of measuring pleiotropic QTL
Gene density/recombination frequency
Introgression model for QTL clustering
Assuming
 many different loci could produce variant alleles
affecting domestication traits (Turner’s largesse of
the genome)
 introgression from wild relatives during fixation
Linked QTL were preferentially fixed (Le
Thierry D’Ennequin et al. 1999 )
Prediction: clustering should be stronger in
outcrossers than selfers, since selfers will not
have been affected by introgression
Wild Relative
Crop
Outcrosser
q
q
q
q
gene flow
q
Q
Q
Q
Q
q
Q
Q
Selfer
q
q
q
q
Q
q
q
Q
Q
Q
Q
Q
Is clustering an artifact of pleiotropy?
A single pleiotropic QTL could be
detected multiple times (i.e. once for
each trait) giving the false appearance
of clustering
Prediction: if we examine a
conservative set of non-pleiotropic
QTLs, we will not see clustering
Easier said than done!
Does clustering reflect gene density?
Gene density and recombination
frequency vary along the genome
QTL could be clustered because they
map to regions with many genes/cM
Suggested for wheat (Peng et al. 2003)
Expected theoretically (Noor et al. 2001)
Gene density and recombination
Noor et al. 2001 Genetics 159, 581
Tests for QTL clustering
Clustering among linkage groups
 2 test of independence
Clustering within linkage groups
Measured by simulation
Randomly assigned same number of QTL
to linkage groups
Measured distance between neighboring
QTL
Reducing pleiotropy
Earliness
Growth habit
Increase in yield
Gigantism
Seed dispersal
Days to flower
Heading date
Fruiting date
Ripening date
Plant height
No. tillers/plant
No. of branches
Average length of
nodes
No. of nodes
Kernel No./spikelet
Kernel No./plant
Kernel No./panicle
Grain yield
Spikelet No./ spike
Spikelet No./panicle
Spikelet density
Spike No./panicle
Cupules/rank
No. of rows of cupules
Fruit number
Fruit yield
Seed size/weight
Panicle length/weight
Spike length/weight
Fruit diameter/weight
/length
Shattering rate
Brittle rachis
Awn length
Full data
set
Reduced
data set
Relationship of clustering to outcrossing rate
Among
Reduced
Within
Crop
Outcrossing rate
Original
Original
Reduced
Rice
<1%
yes
no
Rice
<1%
yes
yes
Rice
<1%
yes
yes
Common bean
1-5%
yes
yes
Tomato
1-5%
yes
yes
no
no
Wheat
1-5%
yes
yes
yes
no
Pepper
12-15%
yes
no
yes
yes
Cowpea
12-15%
yes
no
yes
no
Eggplant
12-15%
yes
no
yes
no
Sunflower
25-40%
yes
yes
no
no
Pearl millet
25-40%
yes
yes
Pearl millet
25-40%
yes
yes
Maize
>40%
no
no
Wild rice
>90%
no
yes
Clustering in an inbreeder: common bean
Non-clustering in an outcrosser: maize
Is clustering of QTLs even specific
to domestication traits?
Crop
Cross type
Rice
DxD
Sunflower D x D
Outcrossing
rate
Clustering
among
Clustering
within
<1%
no
yes
25-40%
no
no
Maize
DxD
>40%
no
yes
Rice
WxD
<1%
yes
yes
Cowpea
WxD
12-15%
yes
yes
Gene density and QTL density
Are QTLs found where transcripts are dense?
 Used rice map of 6591 transcripts (Wu et al. 2002)
 Counted number of markers in 5 cM windows
Results
 Average no. markers/window: 4.41
 Weighted avg no. markers/window for QTL: 3.49
Thus, QTL clusters in rice are not explained
by transcript density
Relationship between clustering
and QTL homology
 Homologous QTL observed in several systems
 Solanaceae (fruit size, shape)
 Beans (seed size)
 Cereals (grain size, daylength sensitivity, shattering)
 Are these QTL underlain by variation at the same genes?
 Implicitly assumed by many in the field
 But it may not be the case
 If
 chromosomal regions have varying propensities to harbor QTL
 these regions are conserved among species
 Some correspondence in the location of QTL among related
species is to be expected
Conclusions
 What is the genetic architecture of domestication?
 High power studies tend to reveal many minor QTL
 Is this due to domestication bottlenecks?
 Recessive domestication alleles of large effect are not the norm
 Where do domestication alleles come from?
 Similarity of dominance spectrum in selfers and outcrossers
suggests many domestication alleles were selected from standing
variation
 Why are QTL clustered?
 Does not appear to be entirely an artifact of pleiotropy
 Lack of effect of mating system argues against introgression as a
major factor
 Appears to reflect inherent differences among regions of the
genome - but not obviously gene density per centimorgan
 A genome structural basis for clustering may contribute to the
pattern of QTL homology
Thanks to
Maria Chacon
Zongli Xu
John Burke (sunflower)
Lizhong Xiong (rice)
Valerie Poncet (pearl millet)
Raymie Porter and Ron
Phillips (wildrice)