6 Oct 2015
October 7, 2015. Last year's massive genomewide association study (GWAS) of schizophrenia identified 108 risk loci, and more are sure to come, as an even larger study is in the works. In the meantime, people are already using the findings: Three recent papers work with single nucleotide polymorphisms (SNPs) identified by GWAS to grapple with genetic risk for schizophrenia.
In a paper published September 22 in Molecular Psychiatry, researchers led by Michael Owen and Michael O'Donovan of Cardiff University in the United Kingdom explore how common risk alleles combine with rare copy number variants (CNVs)—those duplications or deletions of chunks of DNA that impact many genes and which escalate risk far beyond the miniscule effects of common variants. First author Katherine Tansey and colleagues report that people with schizophrenia carrying a CNV also had a higher burden of GWAS-determined risk alleles than did healthy controls. That burden, measured as a polygenic risk score of the total number of high-probability risk alleles weighted by their effect size, was not significantly different from those without a CNV, and suggests that CNV carriers share some of the same biology of risk with those without rare mutations.
The polygenic risk score was also greater in those CNV carriers who had schizophrenia versus CNV carriers who did not have the disorder. This suggests that common risk alleles may combine with CNVs to put risk above a threshold for disease, as proposed by the polygenic threshold model.
One possibility is that CNVs act like a "double or triple scoop of risk alleles," says Francis McMahon of the National Institute of Mental Health in Bethesda, Maryland, who was not involved in the study. "Hard to tell at this stage, but I think this paper does a lot to help move us toward a unified model of risk." (See his full comment below.)
Also in Molecular Psychiatry, a September 1 publication from Denmark reports higher polygenic risk scores in chronic cases of schizophrenia, defined by a high frequency of hospital admissions. Led by Preben Bo Mortensen and Manuel Mattheisen at Aarhus University in Denmark, the researchers were interested in in why some study samples contributing to the landmark GWAS in 2014 had higher polygenic risk scores than others, with prevalence samples that measure the number of people currently with the disorder having higher polygenic risk scores than incidence samples that measure newly diagnosed cases. Hypothesizing that a key difference could be that prevalence samples capture more people with schizophrenia who have been ill enough to need care for a long time, first author Sandra Meier re-analyzed their incidence sample to take multiple hospitalizations into account. This raised the polygenic risk score, indicating that how one recruits samples matters. It also suggests that common risk alleles—whether the collective burden of them or a selected assortment of them—could contain information about the course of the disorder.
A third recent study, published online September 28 in Nature Genetics, uses GWAS-identified SNPs to estimate how genetically related different disorders and traits are to each other. Though this has been done to some extent in previous studies (see SRF related news report and SRF news report), the new study from researchers at the Broad Institute of MIT and Harvard in Cambridge, Massachusetts, introduces a novel method to do these comparisons more comprehensively and quickly, called cross-trait LD Score Regression. This technique makes use of GWAS summary statistics (thus, not requiring individual genotype information) of all SNPs (not just those that reach genomewide significance) in a way that is not biased by overlap in the samples compared. With it, first author Brendan Bulik-Sullivan of the Broad Institute and colleagues replicated the cross-disorder results showing genetic correlations between schizophrenia and bipolar disorder, as well as major depressive disorder, but less correlation with childhood-onset disorders such as autism and attention deficit hyperactivity disorder. New insights included a small positive genetic correlation between schizophrenia and the eating disorder anorexia nervosa and a low correlation between schizophrenia and smoking history.
Though the current batch of schizophrenia-related SNPs are proving themselves useful, there is still a need to complete the catalog of genetic factors involved in psychiatric illnesses, argue Daniel Geschwind and Jonathan Flint in a review article in Science published September 24. "One important justification for the creation of a more complete catalog of genes and mutations is the need to address disease heterogeneity and to understand composite genetic risk in individuals," they write.—Michele Solis.
Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR; ReproGen Consortium; Psychiatric Genomics Consortium; Genetic Consortium for Anorexia Nervosa of the Wellcome Trust Case Control Consortium, Duncan L, Perry JR, Patterson N, Robinson EB, Daly MJ, Price AL, Neale BM. An atlas of genetic correlations across human diseases and traits. Nat Genet. 2015 Sep 28. Abstract
Geschwind DH, Flint J. Genetics and genomics of psychiatric disease. Science. 2015 Sep 25;349(6255):1489-94. Epub 2015 Sep 24. Abstract
Meier SM, Agerbo E, Maier R, Pedersen CB, Lang M, Grove J, Hollegaard MV, Demontis D, Trabjerg BB, Hjorthoj C, Ripke S, Degenhardt F, Nothen MM, Rujescu D, Maier W; MooDS SCZ Consortium, Werge T, Mors O, Hougaard DM, Børglum AD, Wray NR, Rietschel M, Nordentoft M, Mortensen PB, Mattheisen M. High loading of polygenic risk in cases with chronic schizophrenia. Mol Psychiatry. 2015 Sep 1. Abstract
Tansey KE, Rees E, Linden DE, Ripke S, Chambert KD, Moran JL, McCarroll SA, Holmans P, Kirov G, Walters J, Owen MJ, O'Donovan MC. Common alleles contribute to schizophrenia in CNV carriers. Mol Psychiatry. 2015 Sep 22. Abstract