590
Behavior Genes
and consequently are genetically identical.
Fraternal or dizygotic (DZ) twins, on
the other hand, result from two separate
fertilized eggs and so share on average
only 50% of their genes, the same as
full siblings. Hence, if genes contribute
toward predisposition for a phenotype, MZ
twins will tend to be more similar than
DZ twins. If MZ twin pair correlations
or concordances signi±cantly exceed those
observed for DZ twins, we can conclude
a genetic influence for the phenotype,
with environmental variables accounting
for all the difference between MZ twins.
As full sibs share, on average, the same
number of genes as DZ twins, it would
theoretically be possible to substitute them
into such analyses. However, DZ twins are
a better comparison since they are the
same age and are likely to have shared
more similar environments and major life
events. One of the requirements of twin
studies is that correct determination of
zygosity is essential, and misclassi±cation
can result in meaningless results. For
example, if genetic effects are present
and MZ pairs are incorrectly assigned as
DZ, the DZ correlations will be arti±cially
inflated, decreasing the difference between
MZ and DZ correlations and resulting in
an underestimate of heritability, and an
overestimate of the common environment
effect. In practice, zygosity assignment is
frequently based on questionnaires. These
typically ask such questions as whether
the twins are often mistaken one for the
other by friends or if they are as alike as
‘‘two peas in a pod.’’ While genotyping at
several highly polymorphic DNA markers
is a more accurate method of zygosity
testing, it is also far more costly and may
be uneconomical for very large studies
where the approximately 95% accuracy of
questionnaires may be acceptable.
A potential limitation is that twins have
atypical obstetric and perinatal histories
and suffer from birth complications and
low birth weight more frequently than
nontwins do, and consequently, conclu-
sions drawn from twin studies may not be
generalizable to nontwins. For example, on
average, the early cognitive ability of twins
is slightly lower than that of singletons and
twins probably have a slightly elevated risk
of mental retardation. However, in prac-
tice, the long-term impact of this for most
behaviors is limited, and by approximately
±ve years of age, biological, behavioral, and
emotional differences between twins and
nontwins would have diminished.
The classic twin approach depends on
the ‘‘equal environments assumption’’,
that is, MZ twins share their environments
to roughly the same extent as DZ twins.
However, some studies suggest that MZ
twins may share more similar postnatal en-
vironments, such as having more friends
in common, and they may also be per-
ceived and treated more similarly than DZ
twins. The extent to which this may inval-
idate the results of twin studies has been
tested in a number of ways. One approach
is to attempt to measure aspects of environ-
mental sharing and examine whether they
are associated with the trait being studied.
For example, a study of depressive disor-
der in adult same-sex twins found that MZ
twins had on average shared a bedroom for
longer as children and were in more fre-
quent contact as adults than DZ twins, but
these indicators of environmental sharing
did not predict concordance for depres-
sion. Another approach is to look at the ef-
fects of mistaken zygosity on parental treat-
ment of their twins. Up to 20% of parents
think that their MZ twins are DZ or vice
versa. The prediction might be that MZ
twins thought to be DZ by their parents
would be treated less similarly than MZ
previous page 590 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online next page 592 Encyclopedia of Molecular Cell Biology and Molecular Medicine read online Home Toggle text on/off