168
B. F. Voight
heritability remains to be explained by genetics. One intellectual quandary that
has emerged from complex trait mapping studies is that, taken collectively, the
current set of genetic associations does not entirely explain the entire interindividual
predisposition to these traits.
For example, the current estimates for human height suggest that while trait is
upward of 80% heritable, established genetic associations account for only 5–15%
of that heritability (Lango Allen et al. 2010). Similar reports have been shown
for a range of anthropometric, cardiovascular, glycemic, and autoimmune traits
(Manolio et al. 2009). Thus, where does the remaining genetic explanation to this
“missing” heritability reside? A wide variety of potential culprits could explain this
difference: additional risk could reside within genetic variation which is lower frequency, rare, or structural (i.e., deletions or duplications) that has historically been
difficult to assay, excess polygenicity from hundreds of weakly penetrant common
variants which have yet to reach statistical thresholds for significance, epigenetic
phenomenon, parent of origin effects, or within interactions between genes (Maher
2008; Eichler et al. 2010). Another possibility could be that population calculations
might actually overestimate the total heritability, due to simplifying assumptions
about how the traits are modeled with respect to interactions (Zuk et al. 2012). Given
numerous potential sources to explain the unexplained, it is clear that no single
factor will completely equalize the population and genetic estimates of heritability.
7.3.3 The Data Are Consistent with a Strong Polygenic Component
One quite clear source of missing heritability can be explained, in part, by the large
polygenic component underlying complex traits. The evidence for this claim has
been delivered by the development of statistical methods that infer the number
of loci that are left to be found, given the power to discover loci which already
have been seen (Park et al. 2010), predictive models which use large numbers
of common variants in a primary stage to predict disease in a second stage
(International Schizophrenia Consortium et al. 2009), or approaches which directly
seek to estimate the heritability contributed by common variants in large genomewide studies (Yang et al. 2011a). These approaches have resulted in many studies
that quantify the heritability for a range of complex traits. In the example of human
height, while established associations at 180 loci can explain 10% of the variation
in height, common variation overall could explain as much as 45% of the variance
(Yang et al. 2011b; Lango Allen et al. 2010). The results from the application of
all these methods are the summary conclusion that, for each trait, several thousand
genetic loci are linked with common variation tested in GWAS that remain to be
discovered.
B. F. Voight
heritability remains to be explained by genetics. One intellectual quandary that
has emerged from complex trait mapping studies is that, taken collectively, the
current set of genetic associations does not entirely explain the entire interindividual
predisposition to these traits.
For example, the current estimates for human height suggest that while trait is
upward of 80% heritable, established genetic associations account for only 5–15%
of that heritability (Lango Allen et al. 2010). Similar reports have been shown
for a range of anthropometric, cardiovascular, glycemic, and autoimmune traits
(Manolio et al. 2009). Thus, where does the remaining genetic explanation to this
“missing” heritability reside? A wide variety of potential culprits could explain this
difference: additional risk could reside within genetic variation which is lower frequency, rare, or structural (i.e., deletions or duplications) that has historically been
difficult to assay, excess polygenicity from hundreds of weakly penetrant common
variants which have yet to reach statistical thresholds for significance, epigenetic
phenomenon, parent of origin effects, or within interactions between genes (Maher
2008; Eichler et al. 2010). Another possibility could be that population calculations
might actually overestimate the total heritability, due to simplifying assumptions
about how the traits are modeled with respect to interactions (Zuk et al. 2012). Given
numerous potential sources to explain the unexplained, it is clear that no single
factor will completely equalize the population and genetic estimates of heritability.
7.3.3 The Data Are Consistent with a Strong Polygenic Component
One quite clear source of missing heritability can be explained, in part, by the large
polygenic component underlying complex traits. The evidence for this claim has
been delivered by the development of statistical methods that infer the number
of loci that are left to be found, given the power to discover loci which already
have been seen (Park et al. 2010), predictive models which use large numbers
of common variants in a primary stage to predict disease in a second stage
(International Schizophrenia Consortium et al. 2009), or approaches which directly
seek to estimate the heritability contributed by common variants in large genomewide studies (Yang et al. 2011a). These approaches have resulted in many studies
that quantify the heritability for a range of complex traits. In the example of human
height, while established associations at 180 loci can explain 10% of the variation
in height, common variation overall could explain as much as 45% of the variance
(Yang et al. 2011b; Lango Allen et al. 2010). The results from the application of
all these methods are the summary conclusion that, for each trait, several thousand
genetic loci are linked with common variation tested in GWAS that remain to be
discovered.