The binary traits are now assumed to arise from a liability threshold model, under which all individuals have an underlying normally distributed trait, called the liability, and all those whose liability exceeds a fixed threshold will exhibit the trait. Although the liability is not directly observed, this model has several advantages for modelling polygenic effects, including independence of the genetic effects from the trait prevalence, and an elegant linear transformation between effects on liability to corresponding effects on the observed (0/1) trait. This model has recently been elucidated by several authors for studying the quantitative genetics of binary traits in humans, and the reader is referred to their papers for more detailed discussion [21], [24], [30].
