Likelihood ratio tests of association for X-linked genes using a case-parents design
Purpose: Although complex disease genes are known to be located on the X-chromosome, few statistical methods have been developed to test association between genotypes at an X-linked genetic marker and disease phenotype.
Method: We propose a likelihood ratio test for association with a biallelic X-linked marker using a case-parents design. The test is based on nuclear families with a single affected proband and allows additional siblings and missing parental genotypes. Extension to haplotypes is based on assumptions of random mating and multiplicative penetrance.
Results: We investigate the power and type I error rate of our approach using simulated data and apply our method to data from diseases showing linkage to the X chromosome. We find that loss of efficiency with missing parental information can be offset by additional sibling genotype information.
Conclusion: Our fine-mapping approach offers great flexibility for testing different penetrance relationships within and between sexes. In addition, unbiased estimation of marker-related disease risks provides a measure of the degree of association. We provide a Windows/Unix-based software package in C and C++ for public use.