For 20 years we have conducted large scale studies into the molecular genetics and epidemiology of melanoma and its risk factors, particularly nevus density which is more heritable than melanoma itself and therefore a better target for gene-finding. With NCI funds we recently followed up a population-based sample of 1,911 melanoma probands to update survival status and risk factor information;we have obtained DNA from all melanoma cases and a sample of unaffected relatives, and have done likewise in previously studied samples of melanoma cases in children (<14y), adolescents (14-19y) and older men (>50y). We now have DNA samples and risk factor information on 2,783 members of approximately 1,500 melanoma families. We also have >4000 DNA samples from adolescent twins and their family members for whom we have nevus counts and pigmentary phenotypes. From the latter study we recently published linkage to flat nevus density on 9p, 9q, 2p and 8q, replicated in one or more recent studies from UK, France (FR) and Netherlands (NL). To find causal variants under our linkage peaks and also in the rest of the genome we shall conduct a genomewide association scan for nevus density (GWAS). In Stage 1 we shall genotype extremes of the nevus density distribution (300 highest, 300 lowest, one individual per family) from our adolescent twins with 1 million SNPs using Illumina Hap1M arrays (>90% of the genome covered at r2>0.9). Arrays are enriched for SNPs in regions of copy number variation, which is plausibly implicated in nevus etiology. Different Bonferroni corrected alpha levels will be used to assess significance of the ~10,000 SNPs under our four nominated linkage peaks (5 x 10-6) and those in the rest of the genome (1.7 x 10-7). In Stage 2 the top 200 significant hits in linkage regions, and the top 200 in the rest of the genome will be typed in replication samples from the UK, FR and NL (N~3600) and in 1500 melanoma cases and 1500 controls from our familial melanoma study. In Stage 3, the top ~50 SNPs from a combined analysis of Stages 1 and 2 will be typed in the remainder of the adolescent nevus and familial melanoma studies (N=3,200). We shall perform within-family and case-control analyses for association of melanoma risk variables including status, age-at-onset, survival, and severity, with SNPs and environmental risk factors, stratified by familial risk and anatomic site of the melanoma. Similar analyses will be conducted for mediating variables such as nevus density, freckling, and pigmentation. We shall test for gene-gene interactions (epistasis) between SNPs of nevus genes and melanoma risk, and for gene-environment interactions of nevogenesis SNPs with measures of sun exposure, on melanoma risk. Relevance: A high mole count is the strongest risk factor for melanoma and our study has the power to find 80% of genes with effect sizes as small as 1% of variation in moliness. These in turn will be tested to assess their risk for melanoma, both singly and in combinations with other genetic and environmental risk factors. The costs of melanoma to society are considerable. In 1994, there were over 28,000 new cases of melanoma in the USA and a total of 6,680 people died from this disease. Incidence rates in Queensland are almost five times those in the United States.
The aim of our research is to contribute to the alleviation of this disease burden, and it is our assessment that the potential benefits of this research greatly outweigh the costs.
Showing the most recent 10 out of 66 publications
