Two-dimensional genome scan identifies multiple genetic interactions in bipolar affective disorder

Abstract

Background: Bipolar disorder is a highly heritable psychiatric condition, the etiology of which remains largely unknown despite extensive efforts to identify susceptibility genes. Interactions between genes of small individual effect could partially explain the difficulties of traditional one-dimensional approaches to identify genetic risk factors.

Methods: A nonparametric linkage (NPL) analysis of 65 Australian extended pedigrees containing 643 genotyped individuals (of whom 40% were diagnosed with affective disorder) was conducted. Chromosome-by-chromosome correlation analysis of family-specific NPL scores was conducted to detect evidence of genetic interaction. Interaction-specific multipoint NPL and permutation analysis was used to assess linkage interdependence, using family weights derived from the alternative interacting chromosome. Finally, a single nucleotide analysis of each interaction region was conducted using the publicly available genome-wide association, datasets (2933 cases, 2534 controls).

Results: Significant NPL peaks were detected on chromosomes 2q24-33, 7q21-31, and 17q11-25 (Z = 3.12, 3.01, and 2.95 respectively), with four additional suggestive peaks identified. Four robust interchromosomal interaction clusters exceeding Bonferroni correction at alpha = .05 (uncorrected p < 5.38e-07) were detected on 11q23-25-2p15-12, 4q32-35-1p36, 12q23-24-4p16-15, and 20q13-9q21-22. This linkage interdependence was determined significant after permutation analysis (p = .002-.0002). A suggestive interaction was observed in the combined data on 2p14-11q23 (uncorrected p = 5.76E-10, Bonferroni corrected p = .068).

Conclusions: This study indicates a complex interplay between multiple loci underlying bipolar disorder susceptibility, and highlights the continuing usefulness of extended pedigrees in complex genetics. The challenge lies in the identification of specific gene interactions and their biological validation.