Traditionally, pedigree-based estimates of heritability have relied upon experiments involving groups of closely-related individuals. For example, two full-sibs are expected to have half their genome in common, so if a trait is heritable, would be expected them to exhibit more similar phenotypic values than two cousins, who on average share only one-eighth.
Recently, however, a Nature Genetics paper by Yang et. al. has popularised the idea of estimating heritability using samples of "completely unrelated" individuals. Even if a pair of individuals have no recent ancestor, some will share a fraction more genetic variants than others purely by chance ... albeit only one or two percent. By applying the same statistical methods used for pedigree-based approaches, it is possible to obtain meaningful estimates with genomic-based methods. The major difference is the change in scale; the genomic approach essentially compares whether individuals who share, say, an extra two percent of the variants are more similar than, say, those who share one percent. Additionally, rather than focus on estimating heritability, the genomic methods calculate the proportion of variation which can be explained by the genetic variants typed, which arguably is a more informative figure.
Our recent work in the UCL Genetics Institute has been catalysed by the Yang et. al. paper (and a handful of follow-up papers from the Visscher Group in Queensland). We have investigated how it is possible to incorporate assumptions regarding the frequency and effect sizes of causal variants, and how the method can be adapted when some individuals are related. We have examined whether we should be alarmed by the implicit assumptions of additivity, and how the techniques can be carried over from a quantitative to a binary trait (e.g. disease status). Finally, we show the results of applying these techniques to our own Epilesy Study, carried out in association with Imperial College London.
Reference:
Common SNPs explain a large proportion of the heritability for human
height. Nature Genetics 2010. Jian Yang, Beben Benyamin, Brian McEvoy,
Peter Visscher et. al.