Journal of Biometrics & Biostatistics

A Family-based Association Method for Pedigree Including Half-Sib Data

Abstract

Yen-Wei Li and Yi-Ju Li

Family datasets could provide good resources for association studies as an initial investigation or a replication study. The current family-based association tests analyze data only from full sibships of a nuclear family or extended pedigrees of related nuclear families. In order to fully exert all possible information in the family data, we propose a “Pedigrees with Half-sibs Association Test� (PHAST) to accommodate half-siblings if they are available. PHAST adopts the idea of transmission score from the Pedigree Disequilibrium Test (PDT) to construct the test statistic. The difference is that it utilizes the identity-by-descent (IBD) information of the marker between sibling pairs (full or half sibs) to construct an allelic transmission statistic. If parental genotypes are missing, EM algorithm is used to infer parental genotypes and compute transmission scores for all possible scenarios. The computer simulation results suggested that our new method has valid type I error rates under varied family structures. Our method could have more power than PDT and FBAT when the sample size of half-sibs increases, especially the families without parental genotypes. In conclusion, our method can serve as an alternative method of the existing family-based association tests. Furthermore, it can relax the ascertainment criteria for studying late onset diseases since limited siblings are available.