Evidence and Implications for Multiplicative Interactions among Loci Predisposing to Human Common Disease

Abstract

Objective: The aim of this study was to utilize information on monozygotic twin concordance rates and linkage studies results for common diseases to predict the likely mode of interaction between susceptibility loci. Methods: We calculated combinations of allele frequency and genotypic relative risk (GRR) that would generate linkage results typically observed in common human diseases. Given these single locus effects, we calculated the expected monozygotic twin concordance assuming different numbers of loci under different interaction models. Results: We demonstrate that, for disorders like schizophrenia, a purely additive model of interaction among loci is not consistent with the available evidence. Instead there are likely significant multiplicative or stronger interactions. Given these interactions, we show that in a diagnostic test based on a subset of predisposing loci, the marginal increase of predictive value rises with each additional locus that is discovered. Our model was consistent with susceptibility alleles being common or rare. Conclusions: Evidence from monozygotic twin concordance rates and linkage results point to a significant degree of multiplicative interaction among loci.