Age-related macular degeneration--a genome scan in extended families

Abstract

We performed a genomewide scan and genetic linkage analysis, to identify loci associated with age-related macular degeneration (AMD). We collected 70 families, ranging from small nuclear families to extended multigenerational pedigrees and consisting of a total of 344 affected and 217 unaffected members available for genotyping. We performed linkage analyses using parametric and allele-sharing models. We performed the analyses on the complete pedigrees but also subdivided the families into nuclear pedigrees. Finally, to dissect potential genetic factors responsible for differences in disease manifestation, we stratified the sample by two major AMD phenotypes (neovascular AMD and geographic atrophy) and by age of affected family members at the time of our evaluation. We have previously demonstrated linkage between AMD and 1q25-31 in a single large family. In the combined sample, we have detected the following loci with scores exceeding a LOD=2 cutoff under at least one of the models considered: 1q31 (HLOD=2.07 at D1S518), 3p13 (HLOD=2.19 at D3S1304/D3S4545), 4q32 (HLOD=2.66 at D4S2368, for the subset of families with predominantly dry AMD), 9q33 (LODZlr=2.01 at D9S930/D9S934), and 10q26 (HLOD=3.06 at D10S1230). Using correlation analysis, we have found a statistically significant correlation between LOD scores at 3p13 and 10q26, providing evidence for epistatic interactions between the loci and, hence, a complex basis of AMD. Our study has identified new loci that should be considered in future mapping and mutational analyses of AMD and has strengthened the evidence in support of loci suggested by other studies.

Figure  1

Genomewide linkage results of the three most significant analyses (models 1A, 2B, and 3A [for summary, see table 1]). Chromosomes 4 and 10 data include additional analyses (under models 2B-dry and 2C, respectively) that have produced notable results. The candidate regions, on chromosomes 4 and 9, constitute novel loci and have not been reported in previous studies. The remaining regions, on chromosomes 1, 3, and 10, have been also reported by previous and ongoing studies. The genetic locations correspond to the Marshfield genetic map (Broman et al. 1998) and are expressed in Kosambi cM. Note that the HLOD and LOD_Zlr curves represent statistics with different numbers of degrees of freedom (for a brief discussion regarding the comparison of the two statistics, see the “Families and Methods” section).