Search

Transmission distortion refers to deviation from the normal 50:50 transmission of alleles from parents to offspring. Identification of genomic regions which undergo distortion is necessary for the correct interpretation of linkage and association studies, since tests of linkage using affected relative pairs and family based tests of association will yield spurious results in the presence of transmission distortion. With the increasing availability of genome-wide high density SNP data (e.g. from the International HapMap project), identification of these loci is now a real possibility. Here we present an analytical formula which demonstrates that the power to detect transmission distortion is a simple function of the number of heterozygous parents in the sample and the level of distortion at the locus. Our results indicate that whilst it will be possible to identify loci undergoing major levels of distortion using tens or hundreds of trios, large sample sizes in the order of tens of thousands of trios will be necessary to detect minor levels of distortion with appreciable power. The corollary is that genome-wide searches are unlikely to identify loci where the level of distortion is small, although they may serve to identify interesting regions worthy of follow up.

Williams & Blangero (1999) derived closed form expressions for the power of a univariate variance components test of linkage for a variety of pedigree structures. We have extended their results by investigating the effect of including monozygotic twins in the design on the power to detect linkage. Specifically, we determined the power associated with a pedigree of size three, where individuals one and two were monozygotic twins and individual three was a full sibling to the twins. The power of this sampling unit was uniformly greater than the power obtained from a sib-pair under the same genetic model. The reason for this was that addition of a second monozygotic twin provided another estimate of the sibling correlation for the particular IBD class. In addition, when the total heritability of the trait was <50%, the number of individuals that needed to be phenotyped was less than that with sib-pairs alone. However, a pedigree consisting of a monozygotic pair and sibling was never as informative as a sib-trio, presumably because the sib-trio provided information about allele sharing between three individuals, whereas the monozygotic twins and sibling unit only provided one such relationship. We conclude that including a monozygotic twin in the analysis is an economical strategy, since only one twin needs to be genotyped.