Genome-wide association analyses on complex diseases: from single-nucleotide polymorphism to copy numbervariation

Abstract

Complex diseases, unlike Mendialian diseases, are often characterized by genetic heterogeneity and multifactorial inheritance, involving defects in genes from the same or multiple alternative pathways. Many congenital diseases and psychiatric disorders are complex diseases, and incur heavy health care burden on the society. With the advancement in high-throughput genotyping technologies and the availability of the human single nucleotide polymorphism (SNP) catalogue, genome-wide association study (GWAS) has been widely used to investigate the genetic component of complex diseases. Copy number variations (CNV) can also be identified using the data from the same SNP array.

Aiming to identify more disease susceptibility loci for complex diseases, separate GWAS using a case-control design were conducted on anorectal malformations (ARMs) and schizophrenia. ARMs are rare congenital diseases with heterogeneous phenotypes which could probably be explained by the genetic heterogeneity among patients, while schizophrenia is a common psychiatric disorder that is well known for its multigenic inheritance. The GWAS studies on ARM and schizophrenia included 4,369 (patients: N=363; controls: N=4,006) and 1,231 Han Chinese (patients: N=381; controls: N=850) respectively. The two studies were mainly focused on investigating the contribution of rare CNVs to the diseases, involving analyses on global CNV burden, rare CNV association, protein-protein interaction (PPI) network, pathway and chromosomal aberrations. The associations of SNPs with ARMs were also examined. Apart from elucidating the genetic components in these two diseases, a systematic analysis on four CNV detection programs (CNV partition, PennCNV, QuantiSNP and iPattern) was also undertaken. In the study of schizophrenia, a new approach in CNV filtering which was based on latent class analysis was adopted to gather information from multiple CNV prediction programs.

The study of ARMs revealed 79 genes which were disrupted by CNVs in patients only. In particular, a de novo duplication of DKK4 (an antagonist of WNT signaling) was identified, and addition of Dkk4 protein was demonstrated to cause ARMs in mice. Another 10 genes uniquely disrupted in ARMs patients are also related to WNT signaling. Interestingly, this pathway was also significantly inferred by CNV in patients with schizophrenia. A different set of genes related to WNT signaling was disrupted in ARMs patients and patients with schizophrenia. WNT signaling is crucial for the development of multiple parts in the embryo. The contribution of different WNT signaling pathways at different development stages may vary. Apart from the WNT signaling pathway, other genes with biological relevance were also implicated in the two studies through gene-network and pathway analyses. The results from these two GWAS studies support our existing understanding of complex diseases that defects in various interacting genes could contribute to the same disease. In summary, the CNV results from the two studies have demonstrated the genetic heterogeneity nature of these two complex diseases. The findings also uncovered a set of putative disease candidate genes, which can be used as reference materials for future genetic research for ARMs and schizophrenia.