Introduction The genetic susceptibility to melanoma remains the subject of intense research worldwide, focusing on either rare, familial high-penetrance sequence variants, or more common moderate- to low-risk predisposing allele. The number of published data on genetic associations of melanoma has increased immensely over the past years implying a number of putative risk alleles in several genes.
Aim To perform a comprehensive field synopsis of all published genetic associations of melanoma and create an on-line database that collects and curates published results in this area.
Methods We have collected and catalogued all genetic association studies published in the field of melanoma from 1992 until October 2009, using the PubMed and HuGE PubLit searches and the Melanoma Molecular Map Project (mmmp). We included all relevant case/control studies examining the frequency of risk alleles in melanoma patients and excluded studies of rare (minor allele frequency <0.01), highly penetrant gene mutations. We also included data from large-scale association studies and genome-wide association studies (GWAS). Random-effects meta-analyses using allelic contrasts were carried out for all genetic associations having genotype data in four or more independent case/control studies. Nominally significant results of the meta-analyses were graded for strength of epidemiological evidence based on a grading system developed by HuGENet. A publicly available, online database was created containing the results of our field synopsis and meta-analyses results.
Results Our search identified 153 studies in the literature reporting on the association of melanoma risk with 744 polymorphisms in 163 different genes. Apart from 5 recent GWAS, most studies used a candidate gene approach and focused on variants of the MC1R gene (19%), ASIP gene (8%), VDR gene (7%), CDKN2A (7%), and OCA2 gene (6%). We meta-analyzed 78 studies focusing on 37 polymorphisms across 14 different genes. Polymorphisms in several genes (MC1R, ASIP, OCA2, SLC45A2, TYR, VDR and CDKN2A), most of which have been associated with pigmentary traits, showed nominally significant effects with average summary odds ratios of 1.51 (range: 0,361–3,086). Grading of the cumulative evidence by the HuGENet criteria showed that 5 of the significant results (MC1R rs1805007, MC1R rs1805008, MC1R rs1805008 in only Caucasian samples, OCA2 rs1800407 and TYR rs1126809) were characterized as having ‘strong’ epidemiological credibility.
Conclusion Our field synopsis and systematic meta-analysis approach offers a powerful tool to assess the current evidence on the genetic susceptibility of melanoma and to track the most viable candidate genes. It also provides an up-to-date on-line knowledge base of the genetic predisposition of melanoma, a prerequisite for the translation of genetic findings into useful applications for disease prevention.