So far in the literature there has been very little investigation of the performance of methods for phasing in isolated populations. In addition, many GWAS cohorts consist of a range of relatedness between the study individuals. Some cohorts contain mixtures of pedigrees, weakly or cryptically related individuals and more distantly related individuals. Methods for carrying out association studies using related individuals have recently been re-discovered in the literature as a powerful approach, with the additional benefit of implicitly avoiding confounding due to population structure [14]–[16]. In addition, explicit detection of tracts of IBD between pairs of individuals is becoming more widely used for detection of disease genes [17]–[20] and for population genetic analyses [21], [22]. More generally, isolated populations offer promise for interrogating common complex diseases [23]. For many such cohorts phasing will be a first step in performing imputation from a reference panel [11] or as part of an IBD detection analysis, so it is interesting to consider the performance of alternative phasing methods.
