Most GWAS involve independently sampled individuals or families of fixed structure (e.g., parent-offspring trios), which can be analyzed using freely available software, such as PLINK (Purcell et al., 2007). The five different types of four-member families that comprise the MCTFR sample, however, present analytical challenges that cannot be handled efficiently within PLINK or other existing software for GWAS analysis. Consequently, we developed an efficient analytical approach for the MCTFR GWAS data based on a Rapid Feasible Generalized Least Squares (RFGLS) algorithm (X. Li, Basu, Miller, Iacono, & McGue, 2011). Briefly, analysis involves regressing a phenotype on a set of pre-specified covariates and a single SNP genotype (coded 0-1-2), repeated for each SNP in the GWAS. Because of the family structure, the residuals from the regression are non-independently distributed with a variance-covariance structure that depends on family type (e.g., it is not the same for MZ and DZ twin families). In RFGLS, rather than estimate the residual variance-covariance matrix for each family type in each SNP regression, we estimate it only once for each family type based on the model with
