Transcript DNA and Gene Expression

Interplay between Genes and
Environment
Gene Expression
• Evolved to be responsive to intracellular
and extracellular environments
• “Biological index” of environmental
influence
• “Phenotype model”
– Assess environmental influence by its change in
gene expression profiles across genome
Caveats
• Know a lot more about genes than environment
– E.g., from base pairs, on chromosomes, transcription,
etc.
• Questions:
– Where are environmental effects expressed in brain?
– How do environmental effects change with
development?
– How does environment cause individual differences in
behaviour?
– Etc., etc., etc….
Environment
•
•
•
•
Gosh, it’s important
Demonstration through quantitative genetics
Heritability rarely more than 50%
Hence, environmental effects typically
going to account for over 50% of individual
differences
• Three rather important discoveries…
Nonshared Environment
• Environmental influences make children in
same family no more similar than children
in other families
• Rather a surprise for traditional psychology
theories…
Genotype-Environment
Correlation
• People create their own experiences,
partially for genetic reasons
• “Nature of nurture”
Genotype-Environment
Interaction
• Effect of environment can depend on
genetics…
• And, effects of genetics can depend on the
environment.
• Genetic sensitivity to environments
Environment
• Shared environment
– Family resemblance not explained by genetics
• Nonshared (unique) environment
– Variance not explained by heredity or shared
environment
– Includes error of measurement
• Note: shared & nonshared not limited to family
environments
– Experience outside family can be shared or nonshared
by siblings
Testing for Shared Environment
• Direct test
– Resemblance among adoptive relatives
– Zero heredity
Testing for Nonshared
Environment
• Direct test
–
–
–
–
Identical twins
Same genetics
Generally share same environment
So, difference due to nonshared
• But, a conservative estimate
– Twins often share special environments that non-twin
sibs don’t
– Need to account for this when calculating nonshared
effect
Specific Nonshared Environment
• Assess elements of environment specific to
each child
• Not aspects shared by all siblings
• What specific factors make children in same
family so different?
• Remember: factors specific to a child can
also be shared by other child(ren) in family
too
Factor and Behaviour
• So you identify a child-specific factor. Does
it actually relate to behaviour?
– E.g., do a difference in parental treatment really
make a kid schizophrenic?
• Turkheimer & Waldron (2000)
– Overall, specific non-shared environment
factors didn’t account for a lot of the variability
However…
• Plomin, Ashbury & Dunn (2001)
• Each factor doesn’t, but add the specific
factors up and you start to get somewhere
Factor
Variance accounted for in
adjustment, personality, cognitive
Birth order
0.01
Differential parental behaviour
0.02
Differential sib interactions
0.02
Differential peer or teacher interactions
0.05
Several other factors
0.03
Total
0.13
Direction
• Remember correlation is not causation
• Does parent treatment cause the behaviour, or does
the behaviour cause the parental treatment?
• Starting to look like the genetics of a child can
have a role on the behaviour of the parent
• Child appearance
– Berkowitz & Frodi (1979): unrelated adults punished
unattractive children more than attractive children
– Allen et al. (1990): mothers less verbal and more
controlling of their children with congenital facial
anomalies
Other Issues
• Chance
– Random events are another factor contributing
to nonshared environment
• Age differences
– Specific factors in nonshared environment are
going to change markedly across development
So…
• Environmental influences do affect
behaviour
• More individual-by-individual than familyby-family
• Family experiences are important, but these
environmental events will affect different
individuals in different ways
Genotype-Environment Correlation
• Genotype can influence environmental factors’
effects
• Individuals have an active role in selecting,
modifying, and constructing their environment
• “Extended phenotype”
• “Niche construction”
Kendler & Baker (2007)
• 55 independent studies that estimated genetic
influences on an environmental variable
• Seven categories
– General and specific stressful life events, parenting as
reported by child, parenting as reported by parent,
family environment, social support, peer interactions,
and marital quality
• 35 environmental measures from these categories
Results
• Weighted heritability estimates from 7-39% for
individual environmental measures
– Most between 15-35%
• Weighted heritability for all environmental
measures was 27%
• An individual’s genetic influences on his/her
environment account for ~25% of the variance in
the nonshared environment component of
behaviour
Implications
• Strongly support the bi-directional model of person
environment inter-relationships
– Humans actively create important aspects of their social
environment and interpersonal relationships
• Molecular genetics advances promote reductionistic
models of “inside the skin” gene effects
– Can’t give full picture of gene to behaviour pathway
– Must consider the extended phenotype pathways
• Standard heritability estimates can’t distinguish
between inside and outside the skin pathways
– Needs to be addressed given non-trivial role of genotypeenvironment correlation
Three Types of GenotypeEnvironment Correlations
• Passive
– Children passively inherit family environments from their parents that
are correlated with their genotype
– Interactions between genetically related individuals
• Evocative (aka reactive)
– Individuals, due to genotype, evoke reactions from other people
– Between anyone who reacts to individuals due to their genetic
proclivities
• Active
– When individuals select, modify, etc. experiences that are correlated
with their genetic propensities
– Between anyone or anything in the environment
Method 1
• Only usable to detect passive
• See if genetically influenced parental traits
correlate with both the environmental measure and
the children’s trait
• Compare correlations in biological and adoptive
families
– Adoptive parents genetically different from adopted
children
• If correlation greater in biological family, there’s a
passive genotype-environmental effect
Method 2
• For evocative and active
• Compare biological parents’ traits and
adoptive families’ environments
• Biological parents share genes with adopted
away children; adoptive parents react to the
adopted children’s genetic propensities
(partially shared with the biological parents)
Method 3
• Can be used on all three types
• Multivariate genetic analysis of correlation
between an environmental measure and a
trait
• Estimates degree to which genetic effects on
one environmental measure overlap with
other genetic effects on a second measure
In a Nutshell…
• Passive seems most important in childhood
• Evocative and, especially, active increase in
significance with development
• With age, individual does more to direct his
or her own environmental interactions, in
part driven by own unique genotype
Genotype-Environment Interaction
• Genotype-environment correlation: role of
genetics in exposing an individual to
environmental factors
• Genotype-environment interaction: individual’s
susceptibility to specific environments due to
genotype
• Effect of environment on phenotype depends on
genotype, or
• Effect of genotype on phenotype depends on
environment
Possibilities
• G has effect without
effect of E
• E has effect without G
• Both G and E have an
effect
• Both G and E have an
effect and interact
with each other
QT
QT
Low risk
G
High risk
G
QT
Low risk
G
High risk
G
Low risk
G
High risk
G
QT
Low risk
G
High risk
G
QT = phenotypic quantitative trait
G = genetic effects
E = environmental effects
High risk E
Low risk E
Study Designs for Testing G-E
Interaction
• Non-human animal studies
– Have the advantage of being able to manipulate and
control both genotype and environment
• Adoption studies
– Can’t manipulate environment experimentally, but can
take advantage of changed environment via adoption
• Twin studies
– Can use one twin’s phenotype as index of co-twin’s
genetic risk for some trait
– Typically determining if heritability differs in two
environments
General Findings
• Non-human animal models tend to show
interactions, but not consistent
– E.g., maze-bright maze-dull in enriched vs.
restricted environments
• Problem for animal models is that there is
limited opportunity for a rat in a cage to
extend his phenotype…
• Some adoption studies support, but others fail to
find, genotype-environment interactions
• Twin studies generally find some effect
• Not a lot of molecular genetic QTL work on this
yet, but what there is shows effects
– E.g., COMT allele and cannabis use study
• Overall, support for modest genotype-environment
interaction
• A lot of variability in results depending on what
environmental measure is being used
– E.g., cognitive, psychopathology, attitude, personality,
etc.