Transcript Sample collection

Genomic Tools And Strategies
For Studying Hereditary
Conditions In Horses
Sofia Mikko, L.S. Andersson, S. Eriksson, J. Axelsson,
and G. Lindgren
Dept. of Animal Breeding and Genetics, SLU, Uppsala,
Sweden
The Horse Genome Project
http://www.uky.edu/Ag/Horsemap/welcome.html
 Started in Kentucky 1995
 Approx.100 researchers from more than
20 countries
 Meetings are sponsered by Dorothy
Russell Havemeyer Foundation
 Strong collaboration to make a genetic
map, and to sequence the horse genome
 "Twilight" was selected as the
representative horse
 The sequence was publicly available in
January 2007 (http://genome.ucsc.edu/)
"Twilight" - Equus
caballus
Photo courtesy of
NHGRI
The Horse Genome Sequence
http://www.broadinstitute.org/mammals/horse
 EquCab2 is a Whole Genome
Shotgun (WGS) assembly at
6.8X
 The assembly is 2.68 Gb
 The final gene-set comprises
 20,436 protein-coding genes
 4400 pseudogenes (including
retrotranposed genes)
Bravo is a male
Thoroughbred, closely related
to Twilight. DNA from his
blood cells was cut into large
fragments to make a “BAC
Library”.
The Equine Genome – EquCab2
Wade et al., Science vol 326, 2009
 24 individuals from 11 breeds were
sequenced to find SNP:s
 > 1 million distinct SNP:s found
 1 SNP/2000 bp
Hrafnhetta is an
Icelandic horse
mare. She was chosen
for random
sequencing for
comparison to the DNA
sequence of Twlight.
EqCab2 50K SNP chip
Wade et al., Science vol 326, 2009
 54,602 SNP:s were selected for the
EqCab 50K SNP chip
 Majority of SNPs are polymorphic
within breeds
 90% of SNPs are ≤ 110 kb apart
 99.9% of SNPs are ≤ 1000 kb apart
 Illumina will no longer provide this
version of the chip
The Equine Genome – EquCab2
Wade et al., Science vol 326, 2009
 The within breed LD in horse is moderate (100-300 kb)
 Slightly shorter (50-70 kb) across breeds
 Absence of strong bottleneck during breed formation
 Many mares are used to maintain population size
 LD shortest for ancient breeds, and longest for Thoroughbred
The Equine Genome – EquCab2
Wade et al., Science vol 326, 2009
 LD is 5X shorter than in dog, but 5X longer than in human.
 LD in Thoroughbred is comparable to the dog
 Strong conserved synteny between human and horse
SNP density required for genome-wide
mapping in the horse
Wade et al., Science vol 326, 2009
A sampling density of 100,000 SNPs were able to obtain mean maximum
r2 values of >0.5 for tested SNPs in all breeds as well as the across breed
groups.
Power of gene mapping using
the EqCab2 SNP chip
Wade et al., Science vol 326, 2009
Number of SNPs needed to differentiate horse haplotypes for
within-breed gene mapping (by simulation)
Desired mean
r2max
0.7
0.8
0.9
1.0
High LD breed
30,000
100,000
175,000
245,000
Moderate LD breed
155,000
225,000
300,000
370,000
Low LD breed
250,000
320,000
390,000
460,000
Estimated from LD, number of haplotypes within haplotype blocks, and
the polymorphism rate.
Tools and strategies when studying
hereditary conditions
1) Pattern of inheritance
2) Population and family
structure
3) Study design
4) Choice of marker
5) Sample collection
6) DNA preparation
7) Genotyping
8) Analysis of results
9) Functional analysis
Molecular studies of mono-, and
multifactorial traits
 Until now the focus has been on monogenic traits
 SCID, OLWS, HYPP, GBED, JEB, HERDA, PSSM, etc
 Now the focus has shifted towards multigenic traits
 Allergies, Osteochondrosis, Bone spavin, etc
Population and family structure
Related individuals
Unrelated individuals
• Large half-sib groups are often
available in farm animals
• Could be difficult to find in
inbred populations
• Risk of ”deflated ” p-values
• Risk of ”inflated” p-values
• Population stratification
• Less population stratification
QQ-plot
MDS-plot
Study designs
Linkage analysis
Association study
 Half-sib families
 Unrelated individuals
 Traits cosegregating with
chromosome regions
 Cases vs controls
 Large chromosomal areas are
covered
 Smaller chromosomal areas
are covered
Figure from Zhu et al 2008
Choice of DNA marker
SNPs
Microsatellites
 High level of automation
 Low level of automation
 Less bench-time
 Time consuming
 Low cost / marker
 High cost / marker
 Low PIC / marker
 High PIC / marker
 High coverage of the genome
 Lower coverage of the genome
 Bias towards the breed used
 No bias on breed level
for selecting the SNPs
E02 Eq020912_2 308088 Q Score : 5.3 Allele 1 : 97.5 ( M )
2000
1000
0
80
90
100
Sample handling
Sample collection
DNA preparation
 Blood, tissue, hair
 Good DNA quality from blood
samples and ”clean” tissue
 Blood, and tissue samples are
more expensive to collect
 Sample collections/Biobanks
 Breeders and horse owners
are helpful
 DNA in hair samples has lower
quality
 Manual / automatic
 Cases & controls should be
prepared in the same way
Silver Dapple Coat Color
Brunberg et al. BMC Genet 2006
 Autosomal dominant trait
 Only eumelanin is diluted
 Pheomelanin is not affected => chestnut carriers
Candidate genes/Comparative approach
Brunberg et al. BMC Genet 2006
 Candidate genes – dilution genes
 Comparative approach
 Silver in mouse
 Merle in dogs
 Dun in chicken
fading vision
 Fading vision in zebra fish
M-, merle
M-+H-, harlekin
dun
Linkage Analysis of Silver Dapple
Brunberg et al. BMC Genet 2006
 Well characterized family material
 1 heterozygous stallion, 34 offspring, and 29 non-silver dams
 41 microsatellites genotyped
 Linkage analysis using CRIMAP
 Sequencing and SNP analysis to confirm mutation in PMEL17
Multiple Congenital Ocular Anomalies
(MCOA)
Andersson et al. BMC Genet 2008, in prep 2010
 Codominant mode of inheritance
 Heterozygotes have cysts in the eye
 Homozygotes have multiple
abnormalities including cysts, wideopen
eyes, deformed pupils
 Similar disease is present in mice
 Most common in Rocky Mountain
ponies, and silver coat colored horses
”Identical by Decent” (IBD)
mapping of MCOA
Andersson et al. BMC Genet 2008, in prep 2010
 11 genetic markers on ECA6q
 4.9 Mb interval
 The causative mutation must be present within a 421 bp
fragment on ECA6q, in the same region as SILV
 More individuals that are unrelated or from other breeds
could shorten the interval.
 Sequencing next step?
Skeletal Atavism In Shetland Ponies
 Most likely a monogenic trait
 Autosomal recessive
inheritance
 Fully elongated ulna and fibula
 Low prevalence
 Not always reported by
breeders
Photo: Lisa Andersson
 Carriers may remain
undetected
Photo: Göran Dalin
Association Study of Skeletal Atavism
 Old samples from a biobank
 6 affected, 18 carriers, and 24
non-carriers
 Average sample success rate
of 0.974
 Stratification detected
 GWA show no significant
peak, but a few ”small” peaks
are detected
 Homozygosity mapping
 Sequencing
Photo: Göran Dalin
Equine Insect Bite Hypersensitivity
(EIBH, Summer Eczema)
 Most common equine skin disease
 Present in many breeds around the world
 Proteins in the saliva of the biting midges Culicoides
is the main allergen
 Itching dermatitis may cause open wounds,
lichenification, crusts, dandruff and alopecia.
Prevalence and heratibility
study of Equine IBH
Eriksson et al Animal 2007
 Prevalence of 8% in Swedish born Icelandic Horses
 Range 0-30% between different parental half-sib groups.
 Phenotypes were graded in four classes
 Heritability estimated to 14% (40-50% on the
underlying, continuous scale)
GWA study of EIBH
 104 cases, and 105 matched
half-sib controls
 54602 SNPs analyzed
 Genotyping rate was 0.99
 539 SNPs had >10% missing
genotyping
 1014 SNPs were not in HWE
 14651 SNPs with MAF<0.05
 Left 38398 SNPs to be analyzed
 Average spacing between
markers is 59.8 kb (1 bp - 1.3 Mb)
 Average maximum LD (r2)=0.3 at
≈14.4 kb
GWA study of EIBH
 No stratification detected
 Matched cases and controls
 Deflation of p-values due to
family structure
 Conservative phenotype inclusion  No GWA found using allelic
case/control, as well as logistic
for cases
regression.
GWA study of EIBH
 No SNP reached genome wide significance
 With risk ratio of 2 we need > 5X the cases
 Power calculations show that we can only detect loci with very
strong effect (RR>7)
 173,000 markers needed to cover the Icelandic horse genome
Performance Traits In Swedish Trotters
Nei: 0.524
Fst: 0.134
The Standardbred trotter (S)
The North Swedish horse (NS)
Nei: 0.503
Fst: 0.130
Nei: 0.367
Fst: 0.081
The North Swedish trotter (NST)
Population structure
• The STRUCTURE analysis of
the microsatellite dataset
(n=122) could not separate NS
and NST
• “Identity By State” cluster
analysis of the SNP dataset
revealed three breeds
separated into three distinct
clusters.
NS
S
NST
Breed diversity in ECA regions
• Regions of positive or
balancing slection have been
detected by Ewens Watterson
test
• Fst values of the microsatellite
data analysis reveal
chromosome areas where the
NST may be closer related to
S than to NS.
• Fst values on SNP data will be
next analysis
Funding
 Formas
 ATG
 SSH
 Carl Trygger
 Helge Axelsson-Johnsson
 Strömsholm Animal Hospital
 SLU:s Animal Hospital (UDS)
 Mälarkliniken Animal Hospital
Collaborators
•
Swedish Univ of Agricultural Sciences
(SLU):
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
Leif Andersson
Lisa Andersson
Jeanette Axelsson
Hans Broström
Emma Brunberg
Göran Dalin
Björn Ekesten
Susanne Eriksson
Freddy Fikse
Katja Grandinsson
Ingrid Jacobsson
Gabriella Lindgren
Jennifer Meadows
Sofia Mikko
Aneta Ringholm
Kaj Sandberg
Hanna Smedstad
Gunilla Thyreen
•
Norsholms Animal Hospital:
–
•
Östra Greda Research Group:
–
•
Gus Cothran
Rytis Juras
Michigan State Univ.:
–
–
–
•
Marie Sundquist
Texas A&M:
–
–
•
Rebecka Frey
Jessica Eason-Butler
Susan Ewart
David Ramsey
Norwegian School of Veterinary
Science:
–
Knut Röed