Transcript Document

10/07/08
Journal Club
Simone Sanna-Cherchi
OUTLINE
Basis of Admixture Mapping as strategy do identify
disease-associated genes
Papers
Implications of MYH9 association to FSGS and nondiabetic ESRD in African Americans
LINKAGE DISEQUILIBIRUM
• The probability of 2 independent events cooccurring is the product of the individual
probabilities
• LD is defined as the non-random association
of alleles at different loci on the same
chromosome/haplotype
LINKAGE DISEQUILIBIRUM
Allele 1
Allele 1
Marker A
Marker B
Genotype at Marker A allows prediction of genotype at marker B
WHAT CAUSES LD
•
•
•
•
Mutation
Admixture
Selection
Random genetic drift
WHAT CAUSES LD
•
•
•
•
Mutation
Admixture
Selection
Random genetic drift
DECAY OF LD WITH TIME
Simple Decay of LD with time
1
theta=0.01
theta=0.05
theta=0.10
0.9
0.8
0.7
D'
0.6
0.5
0.4
0.3
0.2
0.1
0
0
20
40
60
Number of Generations
80
100
AVERAGE EXTENT OF LD
Half-life of LD: Distance at which average D’
drops below 0.5:
–
–
–
–
US: 60kb; Nigerians: 5kb (Reich et al.)
Sardinia, Finns, US, UK: ~50kb (Eaves et al.)
British ancestry: ~50kb (Abecasis et al.)
Sardinian subisolate: 1-4 Mb (Zavattari et al.)
GENE MAPPING VIA GWAS:
IMPLICATIONS
At least 300,000 markers for GWAS in Caucasians (1 SNP every
10kb)
At least 1 million markers for GWAS in Africans (1 SNP every 3kb)
Statistical issues: multiple test adjustments require large data sets
to identify variants with small effect
Limitations to less common diseases in which sample sizes are in
the order of hundreds of patients (ex FSGS)
ADMIXTURE MAPPING HYPOTHESIS
• Disease allele has different frequencies
in source populations
• For cases, disease allele site will be
enriched for one ancestry
African Americans: Admixture
MS
Prostate
Cancer
ADMIXTURE MAPPING: MALD
Chromosomal
admixture
segments
Smith, Nat Rev Genet 2005
EXPERIMENTALLY distinguish African and
European ancestry
West
WestAfrican
AfricanFrequency
Frequency
100%
100%
The most
informative
~1% of SNPs
provide
powerful
information
about
ancestry
80%
80%
60%
60%
40%
40%
Advantages:
20%
20%
0%
0%
0%
0%
-Cases only
20%
20%
40%
40%
60%
60%
80%
80%
European American Frequency
European American Frequency
100%
100%
-Less than
2000 SNPs
for GWAS
POWER OF MALD ANALYSIS
Patterson, Am J Hum Genet 2004
MAPPING BY SEGMENT ANCESTRY
Patterson, Am J Hum Genet 2004
LINKAGE DISEQUILIBRIUM RESOLUTION
DISEASES
SUITABLE FOR
MALD
HYPOTHESES AND STUDY DESIGN
Cumulative lifetime risk for ESRD is 2.1% in European Americans (EA)
and 7.5% in African Americans (AA)
AA have higher risk for several forms of nephropathy: diabetic,
hypertensive, lupus nephritis, FSGS (4-fold risk) and HIVAN (18- to 50fold risk)
MALD: 1,272 SNPs across the genome in 190 AA FSGS and 222
matched controls
Fine Mapping and identification of MYH9
Extension of association to MYH9 in 241 AA FSGS and 611 controls
Replication in 125 EA FSGS and 221 controls
Extension to 525 AA ESRD (241 hypertensive, 284 diabetic) and 192
controls
SUBJECTS
MALD RESULTS
MYH9 FINE MAPPING
20 MYH9 SNPs:
-
Pronounced difference in allelic differences btw African Yorubans and
Europeans
-
Possible functional significance
-
Strong LD with previously typed SNPs associated to FSGS even after
correction for local ancestry
12 SNPs in 4 neighboring genes as control
GENOTYPIC ASSOCIATION SUGGESTS A RECESSIVE MODEL
ASSOCIATION WITH IDIOPATHIC FSGS, HIVAN AND ESRD
HAPLOTYPE ASSOCIATION
CONCLUSIONS
MALD is a increasingly efficient method to discover genes for complex
traits in admixed populations
MALD require less genotyping (200-500 times less markers compared to
GWAS), thus reducing costs and, most importantly, increasing power
(less correction for multiple testing) and simplifying computation
 Amenable for traits with pronounced differences in risk in the parental
populations and for which large sample sizes are difficult to recruit
Most of the difference in risk of developing FSGS in African Americans
compared to European Americans is due to variants in MYH9
Variants at MYH9 account for the higher risk of non-diabetic ESRD in
African Americans, especially ESRD due to hypertensive
nephrosclerosis
1,354 SNPs for MALD
All ESRD
1,372 cases / 806 controls
DM
703 cases / 806 controls
Non-DM
669 cases / 806 controls
HTN
347 cases / 806 controls
FSGS
87 cases / 806 controls
GN
126 cases / 806 controls
HIV
69 cases / 806 controls
Grey=DM ESRD
Black=non-DM ESRD
Grey= controls
Black= non-DM ESRD
MALD RESULTS
MYH9: nonmuscle myosin heavy chain, class II, isoform A
110 kb, 41 exons, highly conserved among mammals and very similar to
other nonmuscle myosin isoforms
Abundantly expressed in kidney, liver, platelets
Involved in different cellular functions, including cell polarity, architecture
and trafficking
Rare mutations in MYH9 cause several autosomal dominant conditions,
such as May-Hegglin anomaly, Fechtner, Epstein and Sebastian
syndromes, featuring macrothrombocytopenia with variable association
of sensorineural deafness, cataract, neutrophil Dohle-like bodies and
glomerular disease (Alport-like)
MYH9 KO mice show embryonic lethality, while heterozygous are normal
except low penetrance hearing loss
MYH9: nonmuscle myosin heavy chain, class II, isoform A
In the kidney is expressed in the glomerulus (podocytes), peritubular
capillaries and tubules
It takes part of the actin cytoskeleton, regulating contractility and cell
integrity
One patients carrying a MYH9 mutation in whom kidney biopsy was
performed at early kidney involvement showed focal podocyte foot
process effacement and loss of slit diaphragm at the EM
Since Mendelian forms of FSGS are caused by mutations in genes
encoding for proteins of the actin cytoskeleton (ACTN4, CD2AP and
others) it is tempting to speculate that more common variants in MYH9
can predispose to podocyte damage and FSGS and HTN
ADMIXTURE MAPPING MARKERS
•
Theoretically:
–
•
Much less markers sufficient
Practically:
1.
2.
Develop marker set for admixture mapping:
Look for high frequency differences
or
Use many markers, ignore non-informative