Kriging: applying geostatistical techniques to the genetic study of complex diseases

Abstract

Complex diseases often display geographic distribution patterns. Therefore, the integration of genetic and environmental factors using geographic information systems (GIS) and specific statistical analyses that consider the spatial dimension of data greatly assist in the research of their gene-environment interactions (GxE). The objectives of the present work were to assess the application of a geostatistical interpolation technique (kriging) in the study of complex diseases with a distinct heterogeneous geographic distribution and to test its performance as an alternative to conventional genetic imputation methods. Using multiple sclerosis as a case study, kriging demonstrated to be a flexible and valuable tool for integrating information from various sources and at a different spatial resolution into a model that easily allowed to visualize its heterogeneous geographic distribution in Europe and to explore the intertwined interactions between several known genetic and environmental risk factors. Even though the performance of kriging did not surpass the results obtained with current imputation techniques, this pilot study revealed a worse performance of the latter for rare variants in chromosomal regions with a low density of markers.