Browsing by Author "Oikkonen, Jaana"

Genome wide linkage and association methods are used to map genes affecting traits with genetic predisposition. In this thesis, I compare the methods suitable for quantitative trait mapping in complex, extended pedigrees. As a case study, gene-mapping study of musical aptitude is performed with these methods. Linkage analysis methods are developed for family studies. However, only a few methods are suitable for extended families with a quantitative trait. Three linkage programs were successfully applied for such data in this study. These programs are the SOLAR, JPSGCS and KELVIN. All of these three programs are based on different methods and thus, the same calculations are not repeated. SOLAR is based on the variance components method, JPSGCS on a graphical method and KELVIN on the Bayesian method. Association analysis is also difficult to implement for large pedigrees, because it is best suited for case-control data. Fortunately, methods are extended also for family-based studies. Here, a genomic control method was used to correct for the familial relationships. The method evaluates the relatedness from the whole genome data and the association tests are corrected for the relatedness rates. This method was implemented from the GenAbel program. As a case study, these methods were applied to a musical aptitude study. The musical aptitude is here understood as an ability to perceive the melody, harmony and rhythm of music, and to recognize structures in set of sounds. These abilities were tested with Carl Seashore s tests for pitch and time and Kai Karma's test for auditory structuring. The data consists of 107 pedigrees and 93 sporadic subjects, comprising in total of 915 individuals. Each family includes 2 - 50 individuals. These individuals were genotyped with a SNP chip for over 700,00 SNPs. The linkage analyses revealed several promising loci for the musical aptitude. The best result was located in 4q12 and it was found with all of the three linkage programs. Most of the other results could also be identified with multiple programs, but some differences also occurred. However, none of the findings could be discovered with association analysis, probably due to a too small sample size.