doi: 10.1111/j.1469-1809.2005.00243.x Incorporating Serotypes into Family Based Association Studies Using the MFG Test S. L. Minassian 1,2, * , C. G. S. Palmer 3 , J. A. Turunen 4 T. Paunio 4 , J. L ¨ onnqvist 5 , L. Peltonen 4 , J. A. Woodward 6 and J. S. Sinsheimer 2,7,8,† 1 Elan Pharmaceuticals, San Diego, California 2 Department of Biostatistics, University of California, Los Angeles, California 3 Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, California 4 Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland 5 Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland 6 Department of Psychology, University of California, Merced, California 7 Department of Human Genetics, University of California, Los Angeles, California 8 Department of Biomathematics, University of California, Los Angeles, California Summary Family based association tests are widely used to detect genetic effects. The focus of this paper is the maternal-fetal genotype (MFG) incompatibility test, a family based association test which can be used to detect genetic effects that contribute to disease, including alleles in the child that increase disease risk, maternal alleles that increase disease risk in the child, and maternal-fetal genotype incompatibilities. Consideration of incomplete data resulting from using serotypes could expand the power of the MFG test for detecting genetic effects. Serotypes may be all that are available in certain families, or preferred because of convenience or low cost, and thus a modification of the MFG test will allow optimal use of such data. The modified MFG likelihood can accommodate the incomplete data that result from using serotypes rather than the corresponding codominant genotypes. The modified MFG test was evaluated with serotypes and genotypes from families with members affected with schizophrenia. In addition, simulation studies were performed. Results of the data analyses and simulation studies showed that serotypes can be used to augment genotypes within a sample, to increase power to detect effects when the candidate gene produces serotypes. Keywords: gene by environment, gene by gene, maternal-fetal genotype incompatibility, missing data, non- codominant data, rhesus incompatibility, serotypes Introduction Family based association tests are widely used to de- tect genetic effects (see as examples Horvath et al. 2001; Schaid & Sommer, 1993; Spielman & Ewens, 1996; Weinberg et al. 1998; Wilcox et al. 1998). † Corresponding author: Janet Sinsheimer, Ph.D., AV-617 Cen- ter for Health Sciences, Dept of Biomathematics, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095-1766, Fax (310) 825-8685. E-mail: janet@mednet.ucla.edu ∗ Present Affiliation: Elan Pharmaceuticals. Consideration of incomplete data resulting from using non-codominant data, such as serotypes rather than genotypes, would expand the power of these tests for detecting effects; however such data requires substantial modification to the statistical model so that serotypes can be used in the analyses. Without loss of generality, this paper focuses on the maternal-fetal genotype incom- patibility (MFG) test (Sinsheimer et al. 2003), a family based association test that can be used to identify loci that increase risk of disease through the affected child’s geno- type (child allelic effects), the maternal genotype (ma- ternal allelic effects), or an interaction of the mother’s C  2006 The Authors Annals of Human Genetics (2006) 70,541–553 541 Journal compilation C  2006 University College London