August 30, 2013. More than 20 regions of the human genome can be linked to risk for schizophrenia, according to the latest progress report from the ongoing effort to identify common gene variants through genomewide association studies (GWAS). Thirteen of the genomic regions highlighted in the paper posted online August 25 in Nature Genetics represent new loci, write Patrick Sullivan of the University of North Carolina, Chapel Hill, and colleagues.
"We increased the number of loci for schizophrenia by a factor of two or three," Sullivan told SRF. He also pointed to new data that support earlier findings that genes coding for calcium channel subunits and microRNA-137 may be altered in schizophrenia. However, Sullivan and his colleagues point out that the "hits" do not identify specific genes as contributing to schizophrenia, but rather identify areas for closer inspection, as they may contain multiple genes, not to mention DNA that codes for RNAs or has no known function.
Sullivan, who is an advisor to SRF, also focused attention on the modeling analyses described in the paper, which suggest that they are now able, using GWAS data, to show that common genetic variation (i.e., found in more than 5 percent of the population) is responsible for one-half to three-quarters of the heritable risk for schizophrenia.
"Given how controversial GWAS is, it shows what this method can accomplish if we actually wield it correctly in the sense of having enough samples," said Sullivan.
Smorgasbord of genotypes
What is new is the most recent GWAS report is a sample of 5,001 schizophrenia cases and 6,243 controls collected by the study's Swedish collaborators. First author Stephan Ripke of the Broad Institute in Cambridge, Massachusetts, and colleagues combined these subjects with the Psychiatric Genomics Consortium sample already reported on (see SRF related news story) to generate an initial meta-analysis of 13,833 cases and 18,310 controls, and genomewide significant hits in 12 genomic regions.
The authors write that imputation for this study—the ability to make educated predictions about missing genotype information—was improved over early GWAS reports because the researchers were able to use data from the more recent 1000 Genomes Study. Ripke and colleagues thus generated >10,000 hits for some degree of association with schizophrenia, considerably more than HapMap imputation generated (<1,600 hits).
The researchers then took the analysis a step further, digging down into the Swedish/PGC results to pull out single nucleotide polymorphisms (SNPs) that were below the cutoff for genomewide significance but had p values of at least 5 x 10-10. These data were combined with data from an additional 7,413 cases and 19,762 controls from several other European schizophrenia cohorts, bringing the combined analysis to more than 21,000 cases and 38,000 controls. It was from this analysis that Ripke and colleagues found genomewide significant hits at 22 genomic regions.
However, the authors do not devote much attention to specific genes, in the recognition that there are a number of genes, not to mention non-gene-coding DNA, in these regions that may be driving a given positive SNP association.
"Most of these associations are in intronic or intergenic regions. Going from genetic variations to what’s happening at a molecular level is one of the big questions," Sullivan said. "I think it’s possible to work that through systematically, quite carefully, as a number of people are doing."
Instead, the researchers discuss trends they see within the data, including the fact that they replicated several signals near genes coding for calcium channel subunits that have been implicated already in schizophrenia or psychiatric disorders more broadly (see SRF related news story; SRF news story; SRF news story), as well as a hit near the region coding for microRNA-137, which helps control gene expression. And similar to previous studies dating back to linkage analysis, Ripke and colleagues found plenty of association in the tangled web of the major histocompatibility complex on chromosome 6.
Which part of nature?
The nature-nurture debate has generally been settled for most physiological traits, as well as most diseases, as a draw, with both at work. For schizophrenia, the outstanding questions are, How much of nature versus nurture, and what accounts for the nature portion—common genetic variants, rare variants, epigenetic marks, or the previously misnamed "junk DNA"?
With the combined Swedish/PGC data, Ripke and colleagues applied two recently developed analysis methods—genomewide complex trait analysis (GCTA) and applied Bayesian polygenic analysis (ABPA)—to model the contribution of common variants to the estimated 65 percent heritability of schizophrenia. They came up with 52 percent (GCTA) and 78 percent (ABPA), two very different numbers that will require more study but an argument for the importance of common variants, in any case.
"As others have pointed out before, the heritability of schizophrenia isn’t missing but rather hidden, and it’s hidden because of insufficient sample size and because our technology isn’t getting at all the things that it needs to," said Sullivan.—Hakon Heimer.
Ripke S, O'Dushlaine C, Chambert K, Moran JL, Kähler AK, Akterin S, Bergen SE, Collins AL, Crowley JJ, Fromer M, Kim Y, Lee SH, Magnusson PK, Sanchez N, Stahl EA, Williams S, Wray NR, Xia K, Bettella F, Borglum AD, Bulik-Sullivan BK, Cormican P, Craddock N, de Leeuw C, Durmishi N, Gill M, Golimbet V, Hamshere ML, Holmans P, Hougaard DM, Kendler KS, Lin K, Morris DW, Mors O, Mortensen PB, Neale BM, O'Neill FA, Owen MJ, Milovancevic MP, Posthuma D, Powell J, Richards AL, Riley BP, Ruderfer D, Rujescu D, Sigurdsson E, Silagadze T, Smit AB, Stefansson H, Steinberg S, Suvisaari J, Tosato S, Verhage M, Walters JT; Multicenter Genetic Studies of Schizophrenia Consortium, Levinson DF, Gejman PV, Kendler KS, Laurent C, Mowry BJ, O'Donovan MC, Owen MJ, Pulver AE, Riley BP, Schwab SG, Wildenauer DB, Dudbridge F, Holmans P, Shi J, Albus M, Alexander M, Campion D, Cohen D, Dikeos D, Duan J, Eichhammer P, Godard S, Hansen M, Lerer FB, Liang KY, Maier W, Mallet J, Nertney DA, Nestadt G, Norton N, O'Neill FA, Papadimitriou GN, Ribble R, Sanders AR, Silverman JM, Walsh D, Williams NM, Wormley B; Psychosis Endophenotypes International Consortium, Arranz MJ, Bakker S, Bender S, Bramon E, Collier D, Crespo-Facorro B, Hall J, Iyegbe C, Jablensky A, Kahn RS, Kalaydjieva L, Lawrie S, Lewis CM, Lin K, Linszen DH, Mata I, McIntosh A, Murray RM, Ophoff RA, Powell J, Rujescu D, Van Os J, Walshe M, Weisbrod M, Wiersma D; Wellcome Trust Case Control Consortium 2; Management Committee:, Donnelly P, Barroso I, Blackwell JM, Bramon E, Brown MA, Casas JP, Corvin AP, Deloukas P, Duncanson A, Jankowski J, Markus HS, Mathew CG, Palmer CN, Plomin R, Rautanen A, Sawcer SJ, Trembath RC, Viswanathan AC, Wood NW; Data and Analysis Group:, Spencer CC, Band G, Bellenguez C, Freeman C, Hellenthal G, Giannoulatou E, Pirinen M, Pearson RD, Strange A, Su Z, Vukcevic D, Donnelly P; DNA, Genotyping, Data QC and Informatics Group:, Langford C, Hunt SE, Edkins S, Gwilliam R, Blackburn H, Bumpstead SJ, Dronov S, Gillman M, Gray E, Hammond N, Jayakumar A, McCann OT, Liddle J, Potter SC, Ravindrarajah R, Ricketts M, Tashakkori-Ghanbaria A, Waller MJ, Weston P, Widaa S, Whittaker P, Barroso I, Deloukas P; Publications Committee:, Mathew CG, Blackwell JM, Brown MA, Corvin AP, McCarthy MI, Spencer CC, Bramon E, Corvin AP, O'Donovan MC, Stefansson K, Scolnick E, Purcell S, McCarroll SA, Sklar P, Hultman CM, Sullivan PF. Genome-wide association analysis identifies 13 new risk loci for schizophrenia. Nat Genet. 2013 Aug 25. Abstract