HGM2002 Poster Abstracts: 8. Disease Mechanisms
POSTER NO: 469
Power to detect linkage to psychiatric genetic diseases using model free linkage approaches
1Yin Yao, 2BingJian Feng, 3Andrew Collins
The focus of our study is to evaluate the power of detecting a common psychiatric trait composed of two loci, using seven allele sharing statistics among which five are implemented in computer software SimWalk2 (Sobel and Lange, 1996), and two are implemented in GENEHUNTER (Kruglyak et al., 1996). Unlike all previous reports which involve evaluations of the power of allele sharing statistics for a single disease locus model, we used a simulated data set of general pedigrees in which a two-locus disease was segregating (Martinez and Goldin 1989) and evaluated the two implementations of several similar NPL statistics in two computer programs. For eight genetic models considered, we showed that 1) the statistic A in SimWalk 2.8 performs best under a recessive-recessive model (RR) and under a recessive-dominant model RD); 2) the S-all in GENEHUNTER 2.2 performs best under a dominant-dominant (DD) model and a dominant-recessive (DR) model; 3) Sall and Spairs work equally well under and additive model (AD); 4) Sall performs best under three models considering different level of genetic heterogeneity (H10, H25 and H50). We also found that the power of Sall to detect linkage is greater than its equivalent statistic 'E', implemented in Simwalk 2.1. The p values associated with statistic E are consistently lower and the difference becomes more pronounced under the H10 model (10% families are linked to the first locus). To summarize, our simulations show optimism that it is possible to use model free statistical approaches to detect linkage underlying complex psychiatric traits with a sample size such as twenty three-generation pedigrees.
Other abstracts in same session