[Multilocus genetic analysis implicates neurodevelopment and immune system in the etiology of schizophrenia].

Background and Purpose: Schizophrenia is a severe psychiatric disorder of poorly understood etiology, characterized by high heritability, multifactorial inheritance and high heterogeneity. Multilocus associaton methods may reduce the genetic heterogeneity and improve the probability of replication between analyses. The aims of our study were twofold: 1. To analyse genetic risk factors of schizophrenia by using multilocus genetic tests. 2. To assess the replication probability attributable to the various multilocus tests.
Methods: Subjects - Discovery set: case-parent trios of unaffected parents and affected probands with a DSM-IV schizophrenia diagnosis (n=16); replication set: schizophrenia cases and unaffected controls (n=5337). Associations of single nucleotide and indel markers were transferred to gene- and geneset-based associations, furthermore to geneset-enrichment tests and functional annotation cluster analyses in a two-staged designs. Associations with p<0.1 from the discovery set were tested in the replication sample. Familywise p-value correction for multiple comparisons were performed during the replication step.
Results: After correction for multiplicity, no significant association or enrichment were detected for gene-based nor canonical pathway analyses, but significant association of the 14q31 cytoband and enrichments of the 5q31 and Xq13 cytobands were found (p&#95;corr: 0.002, 0.006 and 0.048, respectively). Functional annotation clustering yielded statistically significant enrichment scores for clusters of splicing/alternative splicing, neurodevelopment and embryonic development. Improvements in replication probability were found with increased test complexity (P&#95;rep: 0, 0.015, 0.21).
Conclusion: Our results corroborate the involvement of neurodevelopment, synaptic plasticity and immune mechanisms in the etiology of schizophrenia. Also, our findings indicated improvement of replication probability by using multilocus genetic analyses.