Diapositive 1

1
Introduction to the concept of functional genomics
David Meyre, Associate Professor, McMaster
University (meyred_at_mcmaster.ca) HRM 728 Graduate
Course Genetic Epidemiology October, 24th 2014
Population Genomics Program
2
Introduction to the concept of functional genomics
What Is Functional Genomics? The goal of
functional genomics is to understand the
relationship between an organisms genome and its
phenotype. Functional genomics is a field of
molecular biology that is attempting to make use
of the vast wealth of data produced by genome
sequencing projects to describe genome function.
Functional genomics uses high-throughput
techniques like DNA microarrays, proteomics,
epigenomics, metagenomics, metabolomics and
mutation analysis to describe the function and
interactions of genes.
3
The genomic revolution
Human genome sequence
High-throughput technologies
GENES
Biostatistics Bioinformatics
Large human biobanks
FUNCTIONAL GENOMICS
4
Gene identification approaches
Genome-wide linkage
Candidate gene
GENES
Genome-wide association
Homozygosity mapping
Full exome / genome sequencing
5
Classification of human genetic diseases
OBESITY
Syndromic disease ( lt 0.004)
Monogenic disease ( lt 2 )
Polygenic disease ( 20)
6
Genes and causality
SYNDROMIC / MONOGENIC DISEASE
Beyond co-segregation studies, additional
arguments are needed to demonstrate the causal
role of a mutation in the disease
? functional genomics
7
Genes and causality
POLYGENIC DISEASE (e.g. type 2 diabetes)
Beyond association studies, additional arguments
are needed to demonstrate the causal role of a
variant / gene in the disease
? functional genomics
Sladek et al., Nature 2007
8
Introduction to the concept of functional genomics
I-TRANS-ETHNIC FINE MAPPING APPROACH
9
Trans-ethnic fine mapping approach
.Linkage disequilibrium is the non-random
association of alleles at two or more loci . The
human genome is composed of blocks of linkage
disequilibrium . The extent of linkage
disequilibrium blocks varies according to the
ethnic background
10
Trans-ethnic fine mapping approach
SNP2
SNP3
SNP4
SNP5
SNP1
Icelandic
French
Asian
African
Distance (Kb)
Disease-associated LD block
Causal SNP
11
Trans-ethnic fine mapping approach
. Large-scale resequencing and case control
association studies in Icelandic, Danish, West
African and American African subjects identified
the rs903146 as the likely causal type 2
diabetes-associated SNP
12
Introduction to the concept of functional genomics
II-EVOLUTIONARY GENETICS
13
Evolutionary genetics
Natural selection is the gradual, non-random
process by which biological traits become either
more or less common in a population as a function
of differential reproduction of their bearers. It
is a key mechanism of evolution. The term
"natural selection" was popularized by Charles
Darwin.
Evolutionary genetics (Huxley 1942) -advantageous
mutations have been positively selected in human
populations during recent evolution -disadvantageo
us mutations have been negatively selected in
human populations during recent evolution
14
Evolutionary genetics
THRIFTY GENOTYPE HYPOTHESIS the 'thrifty'
genotype would have been advantageous for
hunter-gatherer populations, especially
child-bearing women, because it would allow them
to fatten more quickly during times of abundance.
Fatter individuals carrying the thrifty genes
would thus better survive times of food scarcity.
? Obesity and type 2 diabetes predisposing
mutations may show evidence of positive signature
of evolution
15
Evolutionary genetics
.The LCT rs4988235 T variant confers lactase
persistence . The LCT rs4988235 T variant is
associated with more milk / dairy products
consumption and increased body mass index . The
LCT rs4988235 T variant has a selective advantage
in milk-producing dairy farming populations and
has been submitted to positive selection in
relation with events of cattle domestication .
The LCT rs4988235 T allele frequency is more
frequent in Northern (MAF 0.7) than in Southern
Europe (MAF 0.1)
16
Evolutionary genetics
LCT rs4988235 T allele frequency in UK
Davey-Smith et al., EJHG 2009
17
Evolutionary genetics
. Genome-wide approaches in diverse ethnic
backgrounds have identified several hundreds of
regions showing recent positive natural
selection . New methods are able to identify
causal variants in regions with positive natural
selection signature . The amino-acid change
Lys109Arg in the LEPR gene is as a causal variant
submitted to positive selection . The Lys109Arg
variant is associated with body mass index
variation
Grossman et al., Science 2010
18
Introduction to the concept of functional genomics
III-GENE VARIANT AND FUNCTION
19
Gene variant and function

• Missense, nonsense, frameshift (indels) coding
  mutations altered protein function
• Intron / exon mutations exon skipping
• Copy Number Variants (CNV) modulation of gene
  expression, haplo-insufficiency
• gene variant in the promoter (Transcription
  Factor Biding Site) change in gene expression
• gene variant in 3UTR altered mRNA stability
• gene variant in microRNAs change in expression
• gene variant in a long-range enhancer change in
  expression of another gene
• gene variant in a CpG methylation site change
  in DNA methylation pattern

How to prove causality between a genetic variant
and a biological effect?
20
In silico prediction studies
Mutations PolyPhen-2 PANTHER SIFT SNAP PMUT
K26E - - - - -
M125I - - - -
T175M
N180S -
Y181H -
G226R -
S325N -
T558A NA - - -
G593R NA
deleterious - neutral
Eight coding non-synonymous mutations in the
PCSK1 gene have been identified in extreme obese
patients the Polyphen-2 software (conservation
of the amino-acid across evolution protein
structure) is 100 concordant with in vitro
studies
Creemers et al., Diabetes 2012
21
In vitro functional studies
68 of non-synonymous mutations found in obese
patients are deleterious (test
alpha-MSH)
Stutzmann et al., Diabetes 2008
22
Intron / exon mutations and exon skipping
. Extreme obesity cosegregates with homozygosity
for a G/A substitution in the splice donor site
of exon 16 of the LEPR gene . The intron / exon
mutation induces skipping of exon 16 and a
truncated inactive leptin receptor
Clement et al., Nature 1998
23
CNVs are highly causal variants in mendelian
diseases

• a 600kb heterozygous deletion (30 genes) on
  chromosome 16p11.2 explains 0.7 of morbid
  hyperphagic obesity and is associated with
  developmental delays
• duplications in the same chromosomal region are
  associated with underweight and eating
  restrictive disorders
• SH2B1, a key modulator of the response to the
  satiety hormone leptin, and a Mendelian
  hyperphagic obesity gene, is located in the
  deleted interval

Walters et al., Nature 2010 Jacquemont et al.,
Nature 2012
24
Gene variation in the promoter and gene
expression
. The -11391 GgtA variant in the promoter of the
ACDC/adiponectin gene is associated with higher
in vitro promoter activity and with higher plasma
adiponectin level in lean and in obese children
Bouatia-Naji et al., Diabetes 2006
25
Gene variation in 3UTR and mRNA stability
.AgtG 1044 TGA SNP is included in the ENPP1 risk
haplotype associated with higher ENPP1 plasma
level and risk of obesity / T2D .AgtG 1044 TGA
forms a linkage disequilibrium block in 3UTR
with AgtC 1092 TGA and CgtT1157 TGA .In HLA
cells transfected with either 3UTR variant or
wild-type cDNA, specific ENPP1 mRNA half-life was
increased for those transfected with 3UTR
variant cDNA (t/24.35 vs. 2.55 h p0.001)
Meyre et al., unpublished
26
Gene variation and long-range enhancer
. The obesity-associated FTO intron 1 region
directly interacts with the promoter of IRX3 gene
(580 Kb downstream of FTO) . The intron 1 SNP in
FTO modulates IRX3 (but not FTO) expression .
Irx3-deficient mice display a leanness phenotype
Smemo et al., Nature 2014
27
Gene variation at a CpG methylation site
. Gene variant rs1421085 in intron 1 of FTO is
the main contributor to polygenic obesity (Dina
et al., Nat Genet 2007) . Gene variant
rs7202116, in full linkage disequilibrium with
rs1421085, creates a CpG methylation site and is
associated with increased methylation of a 7.7 kb
regulatory region within FTO . The 7.7 kb
regulatory region encapsulates a Highly-Conserved
non Coding Element that acts as a long range gene
expression enhancer
Bell et al., PLOS One 2010
28
Introduction to the concept of functional genomics
IV-GENE CANDIDACY
29
Gene candidacy
Sometimes several genes are included in a same
linkage disequilibrium block ? how can we
identify the causal gene(s)?
Sladek et al., Nature 2007
30
Gene candidacy
. Are genes in the disease-associated LD block
involved in syndromic / monogenic forms of the
same disease? -loci associated with polygenic
obesity MC4R, BDNF, POMC, PCSK1, SIM1 -GWAS for
complex traits 20 of the GWAS loci include
genes involved in mendelian disorders for the
same trait . Are genes in the disease-associated
LD block involved in a corresponding phenotype in
animal models (KO, Tg, SiRNA)? -loci associated
with polygenic obesity MC4R, BDNF, POMC, PCSK1,
SIM1, FTO, GIPR, NPC1, SH2B1, TBC1D1, NEGR1 - gt
170 genes induce a phenotype of severe obesity in
genetic mice models . Gene function,
biology -function related to energy metabolism
31
Gene candidacy
In order to find the causal gene in a
disease-associated linkage disequilibrium block,
mRNA expression studies can be useful
(microarrays, RT-PCR) 1-Is the gene expressed
in target tissues for the disease (obesity
brain, adipocytes T2D pancreas)? 2-Is the gene
mRNA expression modulated by the disease status
in a relevant tissue? 3-Is the gene mRNA
expression modulated by the disease-associated
SNP in a relevant tissue?
32
Gene candidacy

• . ORMDL3 is one of the 19 genes located in the
  asthma-associated LD block
• . ORMDL3 is expressed in the lung
• . ORMDL3 mRNA level is modulated by asthma
  disease status in lymphoblastoid cell lines
• . ORMDL3 mRNA level is strongly modulated by the
  asthma-associated SNP in lymphoblastoid cell
  lines
• ORMDL3 is a highly relevant candidate gene at
  this locus

Moffatt et al., Nature 2007
33
Cis versus Trans e-QTLs?
. The polymorphism rs9585056 is associated with
T1D, modulates the expression of the cis-gene
GPR183 and the expression of the IRF7 network
genes
Heinig et al., Nature 2010
34
Gene candidacy
. Combination of expression mRNA and GWAS
studies . 27 genes differentially regulated in
adipose tissue of monozygotic twins discordant
for obesity . Hypothesis driven GWAS analysis
for these 27 genes followed by a replication in a
second independent sample identified a novel
obesity gene F13A1
Naukkarinen et al., PLOS Genet 2010
35
Introduction to the concept of functional genomics
V-STUDY OF ENDOPHENOTYPES
36
Study of endophenotypes
. Rs17782313 near MC4R has been associated with
BMI by GWAS . Deleterious coding mutations in
MC4R are the commonest form of monogenic obesity
with hyperphagia and increased stature . If the
SNP modulates the expression / function of MC4R,
we can predict associations with the same traits
in an appropriate direction . The SNP rs17782313
obesity predisposing allele is associated with
more snacking and overeating and increased
stature ? MC4R is a highly relevant candidate
gene at this locus
Stutzmann et al., Int J Obes 2009
37
Introduction to the concept of functional genomics
I-TRANS-ETHNIC FINE MAPPING APPROACH
II-EVOLUTIONARY GENETICS III-GENE VARIANT AND
FUNCTION IV-GENE CANDIDACY V-STUDY OF
ENDOPHENOTYPES
38
FTO, a good illustration of integrative approach
.Novel variants identified in African
populations .FTO SNP shows evidence of positive
natural selection .The SNP is associated with
different patterns of methylation (demethylase) .
FTO complete deficiency leads to a
polymalformative lethal syndrome in humans . FTO
partial deficiency does not relate to
leanness/obesity in humans . FTO knock-out mice
are lean, FTO transgenic mice are obese . FTO is
highly expressed in hypothalamus and is regulated
by fasting and feeding . FTO SNP is associated
with food intake in humans
39
Ichimura et al., Nature 2012
40
(No Transcript)
41
ANY QUESTIONS?
The French fair-play!