Behavior Genes
593
biological parents of nonadoptees as a test
of whether the environment or rearing has
any effect. A third approach is the cross-
fostering study, which compares the rates
of disorder in adoptees with no biologi-
cal risk but raised by an affected adoptive
parent with the disorder in adoptees with
biological but not environmental risk. This
is the most complicated and practically dif-
Fcult design but it allows the possibility
of testing for the effects of genes, envi-
ronment of rearing, and the combination
of the two. An example was a study of
antisocial behavior by Mednick and col-
leagues who found an increased risk of
having a criminal record in male adoptees
with either a criminal adoptive father or
ac
r
im
ina
lb
io
log
i
ca
lfa
the
rbu
tane
ven
higher rate in those who had both a bi-
ological and an adoptive father who had
a criminal record. However, the disorder
where adoption designs have been used
most extensively is schizophrenia. Some of
the main results from schizophrenia adop-
tion studies are summarized in Table 2.
They show a consistent pattern favoring a
genetic contribution but in contrast with
the study of antisocial behavior just men-
tioned, there is little evidence of a family
environmental effect.
A general criticism of adoption studies is
that adoptees may not be representative of
the population as a whole. ±or example, the
circumstances surrounding adoption may
be particularly stressful placing adoptees at
increased risk of behavioral and psychiatric
disorders. Parents who give up their
children for adoption may do so because
of illness and therefore will tend to
have increased levels of psychopathology.
±inally, adoptive parents are likely to be
unrepresentative in the opposite direction
as adoption agencies will tend to screen
out many ‘‘negative’’ attributes such as
preexisting psychopathology when placing
achild.
A related complication with adoption
studies is selective placement. Some agen-
cies attempt to place adoptees with families
that resemble their biological parents in
terms of social or ethnic background. This
could result in genetic relatives, who have
been independently adopted, sharing a
correlated environment. ±or example, a
tendency toward selective placement has
been suggested in some studies of IQ, al-
though there is little evidence for selective
placement in many other traits. Another
fundamental limitation of adoption stud-
iesistha
tadop
t
ionisbecom
ingrareras
a result of increasing contraception and
abortion, along with the removal of the
stigma of being a single parent.
2.4
Partitioning of the Phenotypic Variance
As described above, observed phenotype
(
V
P
)
can
be
partitioned
into
genetic
(
V
G
)
and environmental (
V
E
) components,
such that:
V
P
=
V
G
+
V
E
(
1
)
±urthermore, the genetic variance can be
partitioned into additive genetic effects
(
V
A
) and dominance deviation effects (
V
D
)
that result from interactions of the alleles
found at a single locus. We may also
encounter epistasis, the interaction or
multiplicative effect of genes at different
loci. In practice, this is not easy to
distinguish from dominance in human
studies.
However,
one
example
of
a
disorder where the pattern of inheritance
would be in keeping with epistasis is
schizophrenia, where the concordance rate
(i.e. the proportion) of relatives and of
index cases who are also affected is around
previous page 593 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online next page 595 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online Home Toggle text on/off