Transcript Association Studies and High-throughput Genotyping Technologies

MEDG 505
Pharmacogenomics
March 17, 2005
A. Brooks-Wilson
Reminder: What is Genomics?
According to http://genomics.ucdavis.edu/what.html:
“Genomics is operationally defined as investigations into the
structure and function of very large numbers of genes
undertaken in a simultaneous fashion”
Pharmacogenetics
• “The study of how genes affect people’s response to
medicines” (NIH)
• A subset of complex genetics for which the traits relate
to drugs
• First observed in 1957
• Part of “personalized medicine”
• 20-95% of variability in drug disposition and effects is
thought to be genetic
• Non-genetic factors: age, interacting medications, organ
function
• Drug absorption, distribution, metabolism, excretion
• >30 families of genes
Pharmacogenetics: Examples
• Drug metabolism genes
• NAT2, isoniazid anti-tuberculosis drug hepatotoxicity
• CYP3A5, many drugs
• Thiopurine S-methyltransferase (TPMT), 6-thioguanine
• Drug targets (receptors)
• B2 Adrenergic Receptor, inhaled B agonists for asthma
• Drug transporters
• P-glycoprotein (ABCB1, MDR1), resistance to antiepileptic drugs
• The examples known today are those that come
closest to simple genetic traits
Potential Consequences
•
•
•
•
•
•
Extended / shortened pharmacological effect
Adverse drug reactions
Lack of pro-drug activation
Increased / decreased effective dose
Metabolism by alternative, deleterious pathways
Exacerbated drug-drug interactions
The Goal of Pharmacogenomics
Picture from Perlegen website: www.perlegen.com
Complex Genetics: Concepts
•
•
•
•
•
•
•
Family studies vs. population studies
Penetrance
Genetic heterogeneity
Linkage vs. association
Haplotypes in family and association studies
Genetic variation, SNPs
Genotyping
Types of Genetic Studies
• Family studies
– multi-generation families
• Association studies
– Case / control (easiest to collect)
Penetrance
• Penetrance = the proportion of carriers who
show the phenotype
• Expressivity = severity of the phenotype
Genetic Heterogeneity
• Locus heterogeneity (what we usually refer
to when we talk about genetic
heterogeneity)
• Allelic heterogeneity
Family Studies Identify Highly
Penetrant Mutations
High penetrance disease
allele(s)
Availability of suitable
families is the limiting
factor
Family studies are
effective for only a
minority of conditions
Association Studies Can Identify
Variants with High or Low Penetrance
• Case / control groups
• Not limited to high penetrance alleles
• Amenable to the study of gene-environment interactions
• A preferred approach for the majority of
complex genetic disorders
Complex Diseases / Phenotypes
•
•
•
•
Multigenic (genetic heterogeneity)
Environmental effects (multiple)
Gene-gene interactions
Gene-environment interactions (for
pharmacogenetic traits: age, alcohol consumption,
hepatitis exposure, etc.)
• Association studies will hold up under these
complications but family-based linkage studies
will not!
Linkage vs. Association
• Linkage is to a locus
– different families can be linked to the same
locus but have different disease alleles
– how to take advantage of this in proving a gene
is responsible for a disease
• Association is with an allele
– done in groups or populations
– the allele arose and was propagated in the
population; the haplotype was degraded by
recombination
Genetic Markers
SNPs:
Substitutions, for example, C / T
Most common type of genetic variation
Ideal for association mapping over short distances
1 SNP every ~ 200 base pairs in a population
1 SNP every ~1000 base pairs between 2 individuals
dbSNP: >10M putative SNPs, > 5M validated SNPs
Microsatellites:
(CA)n or other short repeats
More polymorphic than SNPs
Less common than SNPs
1 polymorphic microsatellite per ~ 100,000 base pairs
Best for linkage mapping over long distances, in families
SNPs
• Single Nucleotide Polymorphisms
• Can also use “Indels”, though some
investigators throw them away!
• Synonymous, non-synonymous SNPs
• Mutation vs. polymorphism vs. variant or
variation
• The 1% definition
SNP Databases
•
•
•
•
•
•
dbSNP (more than just human)
Human Genome Variation Database
At least 11 others!
~ 10 million SNPs with minor allele >1%
~ 7 million SNPs with minor allele >5%
~ 50,000 non-synonymous SNPs in the
human genome
Case / Control Studies
1.
2.
3.
4.
5.
6.
7.
Collect blood samples from patients and controls, with consent
Establish database of clinical and epidemiological data
Select ‘candidate’ genes of interest for each trait
Sequence the candidate genes in a small group of patients
Genotype selected variants in case / control groups
Analyze for association with a phenotype
Analyze for gene-gene and gene-environment interactions
Genetic, Ethical, Legal and Social (GELS) issues investigations
Linkage Disequilibrium
• The difference between the observed
frequency of a haplotype and its expected
frequency if all alleles were segregating
randomly
• For adjacent loci: A,a
B,b
• D = PAB - PA x PB
• D is dependent on allele frequencies
• Other related measures also used
Human haplotype blocks . . .
Ancestral chromosomes
Observed pattern of historical recombination in common haplotypes
Rather than
50 kb
. . . Simplify association studies
Ancestral
chromosomes
A disease-causing
mutation arises
Association with
nearby SNPs
SNP1 SNP2
A
C
A
C
G
T
G
T
A
CA
*
A
C
G
TG
G
T
G
CA
*
A
T
A
TG
G
C
Location of mutation
Gene
LD and Association
• Direct association
– asks about the effect of a variant
– if negative, the gene may still be involved!
• Indirect association
– uses LD
– can be more convincingly negative if
haplotypes are assessed
Haplotype Blocks
•
•
•
•
Became clear in October 2001
87% of the genome is in blocks ~> 30 kb
Not all of the genome is in haplotype blocks!
Average block 22 kb, 11kb in African populations (Gabriel
et al, 2002)
• A few common haplotypes at a given locus in a given
population
• African populations generally have the greatest number of
haplotypes and the shortest haplotype blocks
• Strength of LD and size of blocks varies greatly between
regions
How to Generate Haplotypes
• Haplotyping in families
• Physical determination
– long-range PCR, separation of molecules
– cloning of single molecules
– labor intensive
• Estimate haplotype frequencies
– Expectation Maximization algorithm, others
– generate frequencies for case group, control
group
Tag SNPs
Chromosome copy 1
Chromosome copy 2
Chromosome copy 3
Chromosome copy 4
The HapMap
• Reference map for association studies
• Expected to reduce the number of markers required to
conduct effective genome scans for association
• 270 samples from 4 populations:
–
–
–
–
30 Yoruban trios (Nigeria)
45 unrelated Japanese (Tokyo)
45 unrelated Chinese (Beijing)
30 U.S. trios (CEPH, N/W European ancestry)
• >400,000 markers genotyped in all samples, nearly
1M in CEPH trios
Strategies
• Candidate gene based studies
– hypothesis-driven
– must guess (one of) the right gene(s)!!
– Current state of the art
• Genome scans
– “hypothesis-free”
– scans of ~ 1 million markers are now
possible
SNP Discovery is Still Necessary
• Many have been found by multi-read
sequence mining
• Directed public SNP discovery in certain
sets of genes, e.g.:
– SNP500Cancer
– Environmental Genome Project (EGP)
• Individuals used usually “unaffected”
SNP Discovery
All exons and regulatory regions of each gene
Identify regulatory regions by comparative genomics
Bi-directional sequencing
Denaturing High Performance Liquid Chromatography (DHPLC)
Other methods
1
2
3
PCR Set-up:
Packard Multiprobe II liquid handler
Template aliquotting:
Robbins Hydra
PCR and cycle sequencing: MJ Tetrads
5
4
6
Sequencing: ABI 3700s
Purification of PCR
Products: Agencourt
SNP Discovery: PolyPhred and Consed
PolyPhred: Debbie Nickerson; Consed, Phil Green
Sample Output
GG
GA
AA
Genotyping, Technology
• Determining the allele(s) present in a
particular sample at a particular (SNP)
marker
• Many methods
TaqMan (ABI): Uniplex genotyping
TaqMan
TaqMan Output
Homozygous 1,1
Heterozygous
Homozygous 2,2
MassEXTEND REACTION
Allele 1
Allele 2
Unlabeled Primer (23-mer)
Same Primer (23-mer)
TCT
ACT
+Enzyme
+ddATP
+dCTP/dGTP/dTTP
Extended Primer (26-mer)
Diagram courtesy of Sequenom
Allele 2
Allele 2
Allele 1
EXTEND Primer
TG A
ACT
A
TCT
EXTEND Primer
Allele 1
EXTEND Primer
Extended Primer (24mer)
Sequenom MassARRAY: < 12-plex
*
T
C
*
*
A
C
T
G
*
A
G
*
A
G
Diagram courtesy of Sequenom
Illumina BeadArray System: 1152-plex
• 1152-fold multiplexing
• 0.26 ng of genomic DNA per genotype
• $ 0.05 USD per genotype
Total Internal Reflection
Fiber Cladding
Photons
(out)
Fiber Core
Photons
(in)
Fluorescence
Emission
Excitation Beam
cladding
Illumina BeadArray System
B
A
Decoder Oligo
Decode hyb 1
Decode hyb 2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
T/C
P1’
P2’
P1
P2
A
G
Address’
P3’
PCR with
common primers
P3
/\/\/\/
Address
Decode hyb. 1
Decode hyb. 2
Allele Specific
Extension
Product capture
by hybridization
to array
ParAllele Molecular Inversion Probes:
10,000 Plex
Affymetrix Whole Genome Sampling
Analysis: 500,000-plex
Kennedy et al., 2003
Affymetrix:
Allele-Specific Hybridization
PM = perfect match
MM = mismatch
DNA Pooling Strategies
• Reduce the number of genotypes and genotyping
cost, particularly for whole genome scans
• Pool of case DNAs vs. pool of control DNAs
• DNAs must be mixed in precisely equimolar
proportions in the pools!
• Requires a quantitative genotyping technique
• E.g. 40% in cases vs. 20% in controls
• Verify positives by genotyping individual samples