Transcript QTL mapping - Rice Knowledge Bank

Planning breeding programs
for impact
QTL analysis and
Marker aided selection
Marker aided selection
and QTL analysis
References:
• Kearsey, M.J. and Pooni, H.S. 1996. The genetical
analysis of quantitative traits. Chapter 7
• Bernardo, R. 2002. Breeding for quantitative traits in
plants. Chapters 13 and 14
IRRI: Planning Breeding Programs for Impact
Can anyone describe what
a QTL is?
 A gene or chromosomal region that affects a
quantitative trait
 Must be polymorphic (have allelic variation) to have
an effect in a population
 Must be linked to a polymorphic marker allele to be
detected
IRRI: Planning Breeding Programs for Impact
Mapping quantitative trait loci
(QTL)
QTL = underlying genes controlling quantitative traits
•
Measured with large error effects resulting
•
Result is continuous phenotypic distributions
aa
AA
Phenotypic value
1Leibowitz
et al., 1987
QTL mapping
Example: In progeny derived from cross AA x aa:
• Mean of AA lines is 3100 ± s.e.m
• Mean of aa lines is 2900 ± s.e.m
BUT, AA and aa individuals can’t be visually distinguished
Some AA lines will have low yield due to e’s or other
genes
Some aa lines will have high yield due to e’s or other
genes
IRRI: Planning Breeding Programs for Impact
QTL effect
 Additive effect of a QTL allele = a
 Average value of random lines from a cross between
AA and aa parents = P
Mean of AA lines is P + a
Mean of aa lines is P – a
From previous example, what is
the effect of the QTL (a)?
IRRI: Planning Breeding Programs for Impact
Single-marker analysis
DNA markers can be used to map useful genes
using recombination frequencies of linked genes:
A M
QTL
a
m
Marker
• Markers near QTLs co-segregate with them
• Markers tightly linked to QTL detected by ANOVA
• Most gametes from this F1 = AM or am. If crossover
between marker & QTL, Am & aM gametes will be produced
IRRI: Planning Breeding Programs for Impact
Effect of a marker linked to a QTL
Recombination between M and A is R
In RILs derived from MmAa F1, individuals with MM
marker genotype are made up of 2 QTL genotypes: AA
& aa
- If M and A are tightly linked, most = AA
- If M and A are far apart, as many as half = aa
IRRI: Planning Breeding Programs for Impact
So, the effect of marker M is a function of:
(i)
distance from the QTL
(ii)
size of the QTL effect
IRRI: Planning Breeding Programs for Impact
•
MM lines are easily distinguished from mm lines, but AA
lines can’t be distinguished from aa lines
•
If M and A are linked, average of MM lines will differ from
average of mm lines
•
Size of difference can be between 0 and a, depending on
marker-QTL distance
•
Means of MM and mm recombinant inbred lines
MM = P + a(1-2R)
mm = P – a(1-2R)
R = 2r/(1+2r)
IRRI: Planning Breeding Programs for Impact
QTL mapping with molecular markers
DNA markers used to map useful genes using
recombination frequencies of linked genes:
M1
A M2
m1
a m2
• Markers near QTLs co-segregate with them
• Markers tightly linked to QTL detected by ANOVA
IRRI: Planning Breeding Programs for Impact
QTL mapping strategies
All marker-based mapping experiments have same
basic strategy:
1. Select parents that differ for a trait
2. Screen the two parents for polymorphic marker loci
3. Generate recombinant inbred lines (can use F2derived lines)
4. Phenotype (screen in field)
5. Contrast the mean of the MM and mm lines at every
marker locus
6. Declare QTL where (MM-mm) is greatest
IRRI: Planning Breeding Programs for Impact
Single-marker analysis
1. Select parents that differ for a trait
2. Screen the two parents for polymorphic marker loci
3. Generate recombinant inbred lines (can use F2-derived
lines)
4. Phenotype (screen in field)
5. Do a separate ANOVA on the effect of each marker
6. Declare QTL where F-test is significant
IRRI: Planning Breeding Programs for Impact
QTL mapping strategy:
single-marker analysis
200
QTL?
QTL?
Mean of
MM – mm
lines (kg/ha) 100
• ANOVA done for each marker
• QTL declared if t significant
0
20
40
60
80
Map position (cM)
100
120
Single-marker analysis: example
(taken from Kearsey and Pooni, pp. 137-142)
•
25 RILs produced from an F1 between 2 homzygous
parents
•
Parents differ at marker loci A, B, and C on 1
chromosome:
A-------------B------------------------------C
19 cM
•
51 cM
Lines are evaluated in 4-rep trial
Is there a QTL in this region?
IRRI: Planning Breeding Programs for Impact
Measure of QTL contribution to σP2
 Recall that the simplest QTL model divides the
genotypic effect into a QTL effect (A) and an
effect of all other genes within QTL classes
(G(QTL)):
Y=m + G +
e
= m + G(QTL) + A + e
IRRI: Planning Breeding Programs for Impact
Measure of marker contribution to σP2
Y=m + G +
e
= m + G(M) + M + e
IRRI: Planning Breeding Programs for Impact
Single-marker analysis example
1. Select parents that differ for a trait
2. Screen the two parents for polymorphic marker loci
3. Generate recombinant inbred lines (can use F2derived lines)
4. Phenotype (screen in field)
5. Do a separate ANOVA on the effect of each marker
6. Declare QTL where F-test is significant
IRRI: Planning Breeding Programs for Impact
Single-marker analysis example
F-test for the difference between marker genotype
classes = highly significant at locus B
Therefore, there is a QTL at or near marker B
IRRI: Planning Breeding Programs for Impact
Measure of marker contribution to σP2
Y=m + G +
e
= m + G(M) + M + e
IRRI: Planning Breeding Programs for Impact
Broad-sense heritability for a trial
in which 1 QTL is detected
H =
σ2 G
σ2 P
=
σ2G(QTL) + σ2A
σ2G(QTL) + σ2A + (σ2e /r)
IRRI: Planning Breeding Programs for Impact
R2 is the proportion of σ2P
explained by the QTL A
R2
=
σ2 A
σ2 P
=
σ2 A
σ2G(QTL) + σ2A + (σ2e /r)
IRRI: Planning Breeding Programs for Impact
QTL mapping strategy:
single-marker analysis
Problems with single-marker analysis:
 Not very accurate at assigning QTL position because of
recombination between marker and QTL
 Doing a t-test at every marker results in many false
positives (this is a general problem with QTLs)
IRRI: Planning Breeding Programs for Impact
QTL mapping strategy:
Interval mapping
•
Marker interval = the segment between 2 markers
•
Interval mapping methods use information on values of 2
flanking markers to estimate QTL position
•
The probability that the data could be obtained assuming
a QTL at several positions between the markers is
calculated
•
QTL = declared where the probability of obtaining the
observed data is highest
IRRI: Planning Breeding Programs for Impact
Finding the position of QTL
with molecular markers
DNA markers can be used to map useful genes
using recombination frequencies of linked genes:
M1
A M2
m1
a m2
Recombinant gametes: M1a, m1A,
Parental gametes:
M1A, m1a,
Frequency of recombinants is map distance
What are the problems
with interval mapping?
•
Can’t resolve 2 QTL in a marker interval
•
Although the LOD thresholds seem very high, too many
QTLs are declared (all methods do)
•
Ignores epitasis
•
Not accurate for QTL with small effects (no methods are)
IRRI: Planning Breeding Programs for Impact
Linkage mapping with
molecular markers
Double crossover products look like parental types,
leading to map distance underestimates:
M1
A M2
m1
a m2
Haldane and Kosambi mapping functions used to correct
recombination frequencies
IRRI: Planning Breeding Programs for Impact
QTL mapping strategy:
interval mapping
Significance test:
 Logarithm of the odds ratio (LOD score):
probability of the data occurring with a QTL
Odds ratio =
probability of the data occurring with no QTL
• LOD of 2 means that it is 100x more likely that a QTL exists in the
interval than that there is no QTL
• LOD of 3 means that it is 1000x more likely
IRRI: Planning Breeding Programs for Impact
Fine mapping
•
To be useful in breeding applications, gene of interest
must be tightly linked to marker
•
Ideally, gene itself is used as marker
•
Process of “tagging” gene means it must be cloned
through:
1. Fine-mapping
2. Assigning to a cloned fragment in a DNA library
3. Sequencing
IRRI: Planning Breeding Programs for Impact
Marker-assisted backcrossing
• Main application of gene-tagging is markerassisted backcrossing of recessive genes
• Permits “pyramiding” of resistance genes with
similar phenotypic effects in a screen, e.g Pi1
and Pi2
• Permits rapid recovery of recurrent-parent
genome
IRRI: Planning Breeding Programs for Impact
How is QTL mapping best used?
1. QTL mapping = very inaccurate for detecting,
localizing, and estimating the effect size of genes
with a small effect
2. If repeatability QTL phenotyping experiment = low
QTL map very unreliable
3. QTL mapping works very well to find single genes
with large effects
4. QTL mapping requires a phenotypic screening
system with high H
IRRI: Planning Breeding Programs for Impact
Does anyone else have
other advice on QTL
usage?
IRRI: Planning Breeding Programs for Impact
Some guidelines for
successful QTL mapping
•
Focus on lines that are easy to see in a good screen
•
Derive traits where difference between susceptible and
resistant mapping populations from crosses between
highly resistant and highly susceptible lines
•
Use highly reliable screening systems, and that are
known to differentiate resistant from susceptible lines
•
Do analysis on the means of repeated screens rather
than single trials
•
Ensure that repeatability of your screen is as high as
possible (0.7 or higher)
IRRI: Planning Breeding Programs for Impact
Using QTL in breeding
•
QTLs with small effects = hard to accurately map
•
Only QTLs that are localized to very small chromosome
segments can be successfully used in marker-aided
backcrossing
•
Fine-mapped QTLs with big effects in most genetic
backgrounds and most environments are most useful
e.g. disease resistance genes, Sub1
IRRI: Planning Breeding Programs for Impact
Can anyone briefly explain QTL
mapping strategy?
(single-marker analysis &
interval mapping)
IRRI: Planning Breeding Programs for Impact
Summary
•
QTL mapping = process of locating genes with effects on
quantitative traits using molecular markers
•
QTL mapping strategies = based on measuring the mean
difference between lines with contrasting marker alleles
•
QTL mapping = preliminary step in the discovery of
useful genes for marker-aided backcrossing
IRRI: Planning Breeding Programs for Impact
Summary
•
So far, only successful with disease resistance and
stress tolerance genes having very large effects
•
QTL mapping = basic research activity requiring careful
planning of crosses and high-precision phenotyping
IRRI: Planning Breeding Programs for Impact