Bioinformatics and Approaches to Identifying Polygenic Susceptibility Traits

The role of genetic factors in periodontal disease is now well recognized, although details for the genetic mechanisms of the disease and implications for therapy can be as obscure as they are for other human traits. This paper addresses the role that the analysis of genome-wide data might play in helping to understand the molecular determinants of periodontal risk. Very few human diseases are not polygenic, in that an individual's susceptibility depends on his or her constitution at many genetic loci, each of which may have a small effect. Not only do these loci interact, but also their actions and interactions depend on nongenetic factors. Much of the statistical machinery to handle this complexity was developed in the plant and animal breeding context, where crosses between inbred lines selected for trait differences could be conducted. Human polygenic studies began with studies on large pedigrees, but have expanded to include case-control analyses of random samples of individuals who differ in disease status, and studies of marker transmissions within nuclear families. In the area of characterizing the genetic architecture of complex traits, the relatively new field of bioinformatics is distinguished from the more mature fields of statistical genetics or genetic epidemiology by its focus on genome-wide data. The very dense sets of genetic markers now available, particularly those at single nucleotide positions (SNPs), have meant that it is possible to seek linkages or associations between chromosomal position and disease from the whole genome in a single study. Apart from the obvious problems of scale, there are real issues involved with multiple testing and recognizing interactions. Current thinking tends to focus on relatively conserved “haplotype blocks” instead of single genetic markers, although there is no consensus on the utility of this emphasis. Ann Periodontol 2002;7:1-7.