Population genomics in Sardinia: a novel approach to hunt for genomic combinations underlying complex traits and diseases

The availability of highly polymorphic markers permits testing whether comp lex traits and diseases result from genomic interactions between nonallelic normal variants at separate loci. Such variants may be identified by devia tions from the expected distributions of alleles at a high number of polymo rphic loci, when individuals with the phenotype of interest are compared to normal controls of the same breeding unit, provided that both groups share the same remote ancestry and had no ancestors in common for the last three to four generations. The circumstances needed for such studies are ideally met on the island of Sardinia. The recurrent finding of the same type of a ssociation in separate breeding units between the phenotype of interest and a given genotype should allow a distinction between true genetic identity by descent and randomly occurring identities, as these will be obviously di fferent in separate breeding units. The availability of several breeding un its located in sharply different ecological environments will permit assess ment of the role of nature/nurture factors in the degree of manifestation o f each newly discovered genotype/phenotype association. A pilot study to ev aluate the proposed strategy has been carried out in the Sardinian village of Carloforte, a community of about 8,000 individuals who have remained gen etically homogeneous. Fifty-five control samples have been genotyped with s ix tetranucleotide microsatellites and with a subset of the 400 markers con tained in the ABI PRISM linkage mapping panel, version 2. The allele freque ncies for these microsatellite markers have been determined for these 55 in dividuals and compared to those from a random sampling of subsets of these 55 persons. For the six tetranucleotide microsatellites, a subset of as few as 20 people displayed the same allele frequency distributions as observed with the original 55 unrelated individuals. In conclusion, when samples ar e chosen from the same breeding unit, the number of individuals sufficient to draw the genomic profile of an isolated population can be relatively sma ll. Likewise, the number of probands with the phenotype of interest can be even smaller when they are ascertained with the same genealogical criteria as the normal controls. By comparing the genomic profile of the probands to a fraction of the control samples within each of several separate breeding units of common remote ancestry, the search for genotype/phenotype associa tion for mono- and multifactorial traits and diseases should be simplified and yield unequivocal results.