Transcript Marth-Pfizer-2005-Visit

Computational research
for medical discovery at
Boston College Biology
Gabor T. Marth
Boston College
Department of Biology
[email protected]
http://clavius.bc.edu/marthlab
We study genetic variations because…
… they underlie
phenotypic
differences
… cause heritable diseases
and determine responses
to drugs
… allow tracking ancestral
human history
Our current projects investigate three essential
aspects of genetic variations…
• how to discover inherited genetic polymorphisms that lead
to disease?
• how to model human polymorphism structure to inform
medical research?
• how to select the best genetic markers for clinical
case-control association studies?
1. We build computer tools for variation discovery…
1.
inherited (germ line)
polymorphisms are
important as they can
predispose to disease
the most common type of human
polymorphisms are single-nucleotide
polymorphisms (SNPs) and short
insertion-deletions (INDELs)
P( SNP ) 

all var iable
P( S N | RN )
P( S1 | R1 )
 ... 
 PPr ior ( S1 ,..., S N )
PPr ior ( S1 )
PPr ior ( S N )
P( SiN | R1 )
P( Si1 | R1 )
S
 ... 
 PPr ior ( Si1 ,..., SiN )
 ... 
PPr ior ( SiN )
S i1 [ A ,C ,G ,T ] S iN [ A ,C ,G ,T ] PPr ior ( S i1 )
Marth et al.
Nature Genetics 1999
we have developed a computer package,
PolyBayes© , for accurate discovery of
DNA polymorphisms in clonal sequences
… we are currently expanding our polymorphism detection
capabilities.
Homozygous C
Heterozygous C/T
• for automated detection of somatic single
base pair mutations in diploid samples
Homozygous T
• to include our new knowledge of
human variation structure into the
detection algorithms
• to make the software available for genome centers with
high-performance systems and small Biology labs with
desktop computers
2. We measure genome-wise distributions of DNA
polymorphism data…
0.3
0.2
1. marker density (MD): distribution of
number of SNPs in pairs of sequences
0.1
0
0
1
2
3
4
5
6
7
8
9
10
0.1
0.05
0
1
2
“rare”
3
4
5
6
7
8
9
10
“common”
2. allele frequency spectrum (AFS):
distribution of SNPs according to
allele frequency in a set of samples
… we build models of these distributions under competing
scenarios of human demographic history…
stationary
past
collapse
expansion
bottleneck
history
present
MD
(simulation)
0.3
0.3
0.3
0.3
0.2
0.2
0.2
0.2
0.1
0.1
0.1
0.1
0
0
0
AFS
(direct form)
1
2
3
4
5
6
7
8
9
10
0
0
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
4
5
6
7
8
9
0
10
0.1
0.1
0.1
0.1
0.05
0.05
0.05
0.05
0
0
1
2
3
4
5
6
7
8
9
10
0
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9
10
9
10
0
1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
… and determine the best-fitting models.
European data
African data
genetic
bottleneck
modest but
uninterrupted
expansion
Marth et al.
PNAS 2003; Genetics 2004
3. The HapMap project aims to map out human polymorphism
structure to aid gene mapping…
However, the variation structure
observed in the reference DNA
samples genotyped by the HapMap
project…
… often does not match the structure
in another set of samples such as
clinical samples used to find disease
genes and disease-causing genetic
variants
… we generate “quasi-samples” with computational means to
study sample-to-sample variability…
Instead of genotyping additional
sets of (clinical) samples with
costly experimentation, and
comparing the variation structure
of these consecutive sets directly…
… we generate additional samples
with computational means, based
on our Population Genetic models
of demographic history, using the
Coalescent process.
… and to optimize tag SNP (marker) selection for clinical
association studies.
1. select markers (tag
SNPs) with standard
methods
2. generate computational
samples for this genome
region
3. test the performance of
markers across consecutive
sets of computational
samples
We are developing projects to expand…
• from single-nucleotide DNA changes to developing computer
tools for the detection of other types of genomic and epigenetic
changes (e.g. in cancer)
(Image from Nature
Reviews Genetics)
• to developing visualization and statistical tools for the
integration of diverse genetic and epigenetic data
• to using the fruits of the HapMap project, dense SNPs,
Linkage Disequilibrium, and haplotype markers to help predict
individual responses to drugs, including adverse drug
reactions
Detecting SNPs in medical re-sequencing data, short
insertions / deletions
• detection in new data types produced by the
latest, super-high throughput sequencing
technologies (i.e. 454 Life Sciences sequencing
machines) that will be used for individual
medical re-sequencing
• reliable detection of INDELs and
microsatellite polymorphisms, both in clonal and
in diploid sequence data, e.g. to detect repeat
instabilities
Using SNP array data intelligently to detect chromosomal
aberrations
Speicher & Carter, NRG 2005
Software development for other genetic and epigenetic
data (focus on data confidence)
copy number detection
methylation profile
Laird, NRC 2005
chromatin structure
Sproul, NRG 2005
Integrate genetic and epigenetic data from varied sources
to find “common themes” during cancer development
methylation
profile
chromosome
rearrangements
chromatin
structure
copy number
changes
gene
expression
profile
repeat expansions
Using new haplotype resources to connect genotype and clinical
outcome in pharmaco-genetic systems
• the HapMap was designed as a tool to detect high-frequency (common)
phenotypic (e.g. disease-causing) alleles
• important drug metabolizing enzymes are relatively few in number, well
studied, are at known genome locations, many associated phenotypes are
well described
• many functional alleles are known, and of high frequency (common)
• multi-SNP alleles are highly predictive of metabolic phenotype
• clinical phenotype (adverse drug reaction) less predictable
• ideal candidate for applying haplotype resources
Multi-marker haplotypes as accurate markers for ADRs?
genetic marker (haplotype)
in genome regions of drug
metabolizing enzyme
(DME) genes
computational prediction
based on haplotype
structure
functional allele (known
metabolic polymorphism)
clinical endpoint
(adverse drug reaction)
molecular phenotype (drug
concentration measured in
blood plasma)
Resources
• functional alleles
• LD and haplotype structure in the
HapMap reference samples, based on
high-density SNP map
• specifics of enzymedrug interactions
• existing DME P genotyping chips
Evolutionary / PopGen questions
• mutations single-origin or recurrent?
• geographic origin of mutations?
• mutation age?
• analysis based on complete local
variation structure and haplotype
background of functional mutations
• specifics of the selection process that led
to specific functional alleles?
Proposed steps of analysis
• complete polymorphic structure?
• ethnicity?
haplotype block?
• additional functional SNPs?
• haplotypes vs. functional alleles?
• haplotypes vs. metabolic phenotype?
• haplotypes vs. ADR phenotype?
clinical phenotype
(ADR)
haplotype
functional allele
(genotype)
metabolic phenotype
Funding sources / plans
• polymorphism discovery + medical re-sequencing data
analysis: 5-year NIH R01 research grant awarded
• pop-gen modeling + haplotype analysis + marker selection
system: NIH R01 application pending
• informatics tools for genomic and epigenetic changes in
cancer: need a postdoc to establish project (startup or NIH
R21 or private funding)
• haplotypes in Pharmacogenomics: need a postdoc to
establish project (startup or NIH R21 or private funding)