Transcript Slide 1

The Genetics of Complex Traits:
What Are the Implications for Education?
Michael J. Dougherty
Director of Education, ASHG
[email protected]
NCHPEG Meeting, September 23, 2009
How are U.S. students doing
in science and math?
TIMSS, 2008
General trend in
U.S. science
achievement
from fourth to
twelfth grade is
downward.
“. . . there is actually a
steady decline, not a
sudden drop, in
performance as
students progress
through school . . .”
American Institutes for
Research, 2007
How does genetics education fare in America?
TABLE 1
NAEP test results in 2000 for science reveal a deficit in student understanding of
core genetics concepts (O’Sullivan et al. 2003)
Theme
Classification
Theory of evolution
Reproduction
Evolutionary relationships
Darwin’s theory of evolution
Genes*
Mutation*
Interpreting genetic material*
Genetic disease*
Recombinant DNA usage*
Grade
Students with
complete/essential
answers (%)
Students
with partial
answer (%)
Students with
unsatisfactory
answer (%)
8
8
12
12
12
12
12
12
12
12
23
53
61
25
51
21
2
1
5
8
16
NA
NA
NA
NA
45
33
1
31
27
58
45
39
70
47
30
58
83
56
58
Percentages may not total to 100 due to rounding and student omission (i.e., no answer was given).
*These questions are in the molecular and human genetics category.
Prevalence of misconceptions by genetics topic
500 essays chosen at random (20% of total submitted); systematically
reviewed for misconceptions. From Shaw, et al, Genetics 178: 1157, 2008.
55.6% had one misconception
20.2% had two or more
Percent
Percent of Misconceptions in Sample by Topic
25
20
15
14
17.2
12.8
12.8
10
5
0
Patterns of
Inheritance
Deterministic
Nature of
Genes
Genetic
Technology
Nature of
Genes &
Genetic
Materials
Topic
8.4
8.2
7
Genetic Basis
of Disease
Genetics
Research
Reproductive
Technology
What does recent research indicate about single-gene
disorders and what do we teach?
Phenylketonuria (PKU)
Mutations in phenylalanine hydroxylase gene (PAH) can cause
phenylketonuria. Autosomal recessive.
-Phenotypically there are at least four clinical phenotypes: classic PKU,
moderate, mild, mild hyperphenylalaninemia
Israeli study, n=180, Bercovich et al, J. Hum. Genet., 2008
- Only 63% of metabolic phenotypes could be predicted from exonic
mutations.
- Phenotypic inconsistencies existed when PAH had more than one
mutation.
Ataxia telangiectasia
Autosomal recessive; ATM; rare neurodegenerative disorder
(Smirnov and Cheung, Am J Hum Genet, 2008)
Recessive expression
Carriers
resemble
noncarriers.
Dominant expression
Carriers
resemble
AT patients.
i.e., ATM
genotype is
associated
with the
differential
expression of
other genes.
Why is this complexity important?
May help explain why:
- ATM mutations are susceptibility alleles for breast cancer.
- Carriers have increased risk of heart disease (pleiotropy or
polygeny?).
- Heterozygous ATM mice have increased risk of cancer and
metabolic syndrome.
How common is this variation in “monogenic trait”
phenotype likely to be?
CF- carriers have increased risk of asthma; 1500 mutations; variable
symptom severity.
Gaucher- tremendous symptom variation in patients; carriers have
increased risk of parkinsonism.
FFI and CJD- (autosomal dominant) second-site mutations in PRNP were
thought to be determinative of phenotype; however, variation in pathology
and clinical manifestation implicate other factors.
Do we teach single-gene traits and
disorders in all their complexity?
What is the most common tool used to teach genetics?
From http://anthro.palomar.edu/mendel/mendel_2.htm
Should this be the first and most common
heuristic we use with our students?
“If everyone on both sides of your family is tall,
you are going to be tall. If half are tall and half are
short, you have a 50/50 chance of being either tall
or short. You also have the possibility of ending up
somewhere in the middle.”
If monogenic disease shows such complexity, how
should we think about hypertension, diabetes, weight?
Consider the difference in phenotype distribution
between single-gene and complex traits.
What causes such a distribution?
Genes
+
Environment
Are such traits
common?
Establishing population references for blood glucose and lipids
(Bismenya, et al Afr Health Sci. 2006 December; 6(4): 247–253)
Study:
183 university
students
Blood samples
analyzed
Results:
Near-Gaussian
distributions for
glucose, total
cholesterol, HDL,
and LDL
What do these results suggest about the underlying genetics
and how can we study such traits?
Genome-wide association study (or GWAS)
Compare phenotypic “cases” to “controls” by comparing genetic
differences at many different loci called SNPs.
- Essentially a hypothesis generation technique made possible by highthroughput sequencing technology and massive computing power.
Principle of a DNA
microarray
(from Carr et al, Comp.
Biochem. Physiol., 2008)
http://www.mun.ca/biology/scarr/DNA_Chips.html
What are SNPs?
Single-nucleotide polymorphisms
GWA study of height (Weedon et al, Nature Genetics, 2008)
Red dots represent SNPs with P < 5 x 10-7 in stage 1 and 2 analysis.
What important educational messages
are emerging from GWA studies?
In Weedon study:
- GWAS and genotyping on 30,000 individuals led to 20 variants
associated with height.
-Together, the 20 SNPs explain 3% of height variation.
6%
6%
≤ 17 “tall” alleles
6%
≤ 17 “tall” alleles
5 cm
≥ 27 “tall” alleles
6%
≥ 27 “tall” alleles
As of mid-2008, there were a total of 54 “validated” SNP variants for
height (Visscher). Implicated genes: signaling, ECM, cancer.
What effect would learning through these
examples have on students’ understanding of
patterns of inheritance and the nature of genes?
How might genetics principles related to
complex traits contribute to a more informed
public and better health care?
Genetic variation and common disease (Adeyemo and Rotimi, Public
Health Genomics, 2009):
Used GWAS to investigate loci showing strong and consistent association with
several common diseases.
- Studied 11 populations; estimated allele freq. using HapMap data
Conclusions:
a) wide variation in allele frequencies across populations, up to
40-fold
b) FST, a measure of population differentiation due to drift, varied
widely for the susceptibility loci:
0.12- HapMap autosomal avg. (0 would mean no genetic
substructure, random mating, etc)
0.019 - .201 for type 2 diabetes
0.022 - .520 for prostate cancer
Health risk estimated from any “of these risk alleles is likely to show
wide variation across populations simply as a function of its frequency,
and this risk difference may be amplified by gene-gene and geneenvironment interactions.”
How are the results of GWAS likely to be most useful?
If we focus on genetics alone, can we explain
continuous distributions?
Will students understand this message?
The environment is,
of course, crucial to
complex trait
expression, as it is
for single-gene
disorders.
Some genotypes may “predispose” to certain conditions only in the
presence of environmental triggers. Genetic polymorphism affecting severe
depression. [Data from A. Caspi, et al. 2003. Science 274: 1527.]
Note: A recent meta-analysis by Risch, et al. (JAMA, 2009, 301(23): 2462-71) supports a
link between stress and depression but found no link between serotonin transporter
genotype, alone or in combination with stress.
Environmental influence on phenotype: Do we teach it?
Classic experiments
After 100 generations of inbreeding, mice continue to show metabolic
variability (Storrs and Williams, PNAS, 1968).
The variability in growth of plant clones at different elevations
(Clausen, Keck, and Hiesey, 1940s and 1950s)
High school curriculum
Five out of six leading high school texts lack a sound discussion of
incomplete penetrance (and most fail to distinguish it from variable
expressivity), let alone genetic and environmental modifiers
contributing to it (e.g., the influence of pregnancy and smoking on
carriers of BRCA1/2).
- Also, missing from a majority: “gene-environment interactions and
the potential influence of these interactions on disease” (from L. Doyle,
2009)
An alternative genetics
education paradigm:
Inverting the curriculum
Many phenotypes related to health
and disease vary continuously.
Such traits often have multifactorial
causation; multiple genes and
environmental interactions are
critical.
Additive models of gene action help
explain continuous variation, but
predicting phenotypes from
genotypes is difficult.
Some phenotypes are strongly
influenced by single genes, and
mendelian segregation allows more
precise prediction of phenotypes from
genotypes.
Simplified Conceptual Flow for an Inverted Genetics Curriculum
1.
2.
3.
Many traits show continuous variation (e.g., height, weight, forearm length, extroversion, etc).
Such traits (quantitative/complex traits) can be inherited and are strongly influenced by the
environment.
The level of a quantitative trait can be understood in terms of “contributing factors” that offspring
receive from parents.
a.
many contributing factors (in an additive model) lead to greater manifestation of a trait;
b.
fewer factors lead to less manifestation; and
c.
most combinations lead to an intermediate level of manifestation.
Concepts 4-8 are part of a traditional genetics unit:
4.
5.
6.
7.
8.
9.
Contributing factors that offspring receive from parents are called genes and are carried on the
chromosomes passed from parents to children;
Genes exist in different forms called alleles;
Alleles are passed from generation to generation through the processes of meiosis and fertilization;
The movement of chromosomes (and the alleles they carry) during meiosis and fertilization lead to
characteristic patterns of inheritance;
Following the inheritance of one gene (one pair of alleles) or two genes (two different pairs of
alleles) reveals the patterns of inheritance first identified by Mendel:
a.
monohybrid crosses result in a 3:1 phenotypic ratio and reveal segregation of alleles; and
b.
dihybrid crosses result in a 9:3:3:1 phenotypic ratio and reveal independent assortment of
genes.
The genes and alleles contributing to complex traits segregate and assort according to the same
patterns identified by Mendel except that:
a.
in complex traits, many genes and alleles contribute to one trait rather than each gene
contributing to a separate single-gene trait (e.g., as in dihybrid crosses); and
b.
when only one gene or allele primarily determines a trait (e.g., smooth pea texture vs.
wrinkled), the resulting trait shows a rare pattern of variation (i.e., discrete, not
continuous).
More env influence
Less env influence
Trauma
Height, weight, CVD
CF, HD, AT
Language spoken
------Cancer------
Schizophrenia
Increasing genetic influence
A Continuum of Genetic Influence