Abstract
Epidemiological studies indicate that many immune diseases occur at different rates among people with schizophrenia compared to the general population. Here, we evaluated whether this phenotypic correlation between immune diseases and schizophrenia might be explained by shared genetic risk factors (genetic correlation). We used data from a large genome-wide association study (GWAS) of schizophrenia (N=35,476 cases and 46,839 controls) to compare the genetic architecture of schizophrenia to 19 immune diseases. First, we evaluated the association with schizophrenia of 581 variants previously reported to be associated with immune diseases at genome-wide significance. We identified three variants with pleiotropic effects, located in regions associated with both schizophrenia and immune disease. Our analyses provided the strongest evidence of pleiotropy at rs1734907 (∼85kb upstream of EPHB4), a variant which was associated with increased risk of both Crohn’s disease (OR = 1.16, P = 1.67×10−13) and schizophrenia (OR = 1.07, P = 7.55×10−6). Next, we investigated genome-wide sharing of common variants between schizophrenia and immune diseases using polygenic risk scores (PRS) and cross-trait LD Score regression (LDSC). PRS revealed significant genetic overlap with schizophrenia for narcolepsy (p=4.1×10−4), primary biliary cirrhosis (p=1.4×10−8), psoriasis (p=3.6×10−5), systemic lupus erythematosus (p=2.2×10−8), and ulcerative colitis (p=4.3×10−4). Genetic correlations between these immune diseases and schizophrenia, estimated using LDSC, ranged from 0.10 to 0.18 and were consistent with the expected phenotypic correlation based on epidemiological data. We also observed suggestive evidence of sex-dependent genetic correlation between schizophrenia and multiple sclerosis (interaction p=0.02), with genetic risk scores for multiple sclerosis associated with greater risk of schizophrenia among males but not females. Our findings suggest that shared genetic risk factors contribute to the epidemiological co-occurrence of schizophrenia and certain immune diseases, and suggest that in some cases this genetic correlation is sex-dependent.
Author Summary Immune diseases occur at different rates among patients with schizophrenia compared to the general population. While the reasons for this phenotypic correlation are unclear, shared genetic risk (genetic correlation) has been proposed as a contributing factor. Prior studies have estimated the genetic correlation between schizophrenia and a handful of immune diseases, with conflicting results. Here, we performed a comprehensive cross-disorder investigation of schizophrenia and 19 immune diseases. We identified three individual genetic variants associated with both schizophrenia and immune diseases, including a variant near EPHB4 – a gene whose protein product guides the migration of lymphocytes towards infected cells in the immune system and the migration of neuronal axons in the brain. We demonstrated significant genome-wide genetic correlation between schizophrenia and narcolepsy, primary biliary cirrhosis, psoriasis, systemic lupus erythematosus, and ulcerative colitis. Finally, we identified a potential sex-dependent pleiotropic effect between schizophrenia and multiple sclerosis. Our findings point to shared genetic risk for schizophrenia and at least a subset of immune diseases, which likely contributes to their epidemiological co-occurrence. These results raise the possibility that the same genetic variants may exert their effects on neurons or immune cells to influence the development of psychiatric and immune disorders, respectively.
Introduction
Despite recent advances in identifying key biomarkers and genetic loci for schizophrenia, its pathophysiology remains poorly understood [1, 2]. One interesting epidemiological observation is that the risk of developing many immune-mediated diseases is increased among patients with schizophrenia [3–5], and vice versa [6, 7]. Here, we use the term immune disease to broadly encompass both autoimmune and inflammatory disorders. While there are discrepancies among studies regarding which immune diseases are most strongly correlated with schizophrenia, there is converging evidence that these diseases co-occur at a greater rate than is expected by chance [3–7]. A notable exception is rheumatoid arthritis (RA), where a consistent inverse association with schizophrenia has been observed [5, 8].
Genetic factors have long been proposed as an explanation for the differing prevalence of immune diseases among patients with schizophrenia compared to the general population [5, 6]. The recently reported role of complement component 4 (C4) variation in schizophrenia [9] illustrates a potential shared genetic mechanism in the development of immune and psychiatric disorders. Genetic variants conferring increased C4 expression protect against developing systemic lupus erythematosus (SLE), possibly by increased tagging of apoptotic cells – which are the trigger for autoantibody development in SLE – leading to more effective clearance by macrophages [10]. The same genetic mechanism may increase the risk of developing schizophrenia, by increased tagging of neuronal synapses for elimination by microglia leading to excessive synaptic pruning [9]. We hypothesize that similar shared genetic mechanisms may occur throughout the genome, with cellular manifestations in immune cells and neurons influencing the development of immune and psychiatric disorders, respectively. Previously, we found that susceptibility to schizophrenia does not appear to be driven by the broad set of loci harboring immune genes [11]. However, not all genetic variants conferring risk of immune disease fall within immune loci. Here, we evaluated whether common genetic variants influencing the risk of 19 different immune diseases may also be involved in schizophrenia.
Our cross-disorder genetic approach is supported by recent successes in identifying shared genetic risk variants (pleiotropy) across a variety of human diseases [12–18]. Pleiotropy is emerging as a pervasive phenomenon in the human genome [19–21], and cross-disorder studies characterizing the nature of genotype-phenotype relationships have the potential to yield significant insights into disease etiology. For instance, cross-trait genetic analyses have shed new light on cardiovascular disease and lipid biology – and shifted attention away from HDL as a potential treatment target – by demonstrating that increased HDL cholesterol levels do not reduce the risk of myocardial infarction [14]. In psychiatry, cross-disorder analyses have identified significant pleiotropy between schizophrenia, bipolar disorder, and major depressive disorder, indicating that these diseases are not as distinct at a pathophysiological level as current diagnostic criteria suggest [12, 13, 22].
While previous studies have investigated genome-wide pleiotropy between schizophrenia and immune disorders, results have been inconsistent (S1 Table). Genetic correlation has been reported between schizophrenia and Crohn’s disease [23–27], multiple sclerosis [28], primary biliary cirrhosis [25], psoriasis [29], rheumatoid arthritis [23, 24], systemic lupus erythematosus [24, 25], and type 1 diabetes [23, 24, 26, 27] in some studies, but not in others [8, 13, 16, 24, 30]. Interestingly, negative genetic correlation (whereby genetic risk protects against developing schizophrenia) has also been reported for RA [31], in keeping with the inverse epidemiological association [5, 8].
Additional studies are needed to reconcile the inconsistencies in existing cross-trait analyses of schizophrenia and immune disorders, with careful attention towards potential confounding variables (e.g. population stratification, linkage disequilibrium, non-independence of genome-wide association study (GWAS) samples, and sex-specific effects). To this end we have performed a comprehensive cross-disorder analysis of schizophrenia and 19 immune diseases, using data from the largest available genetic studies. Our findings add to a growing body of literature supporting pervasive pleiotropy between schizophrenia and immune diseases. We extend existing literature by including 10 immune diseases that have not previously been compared with schizophrenia, prioritizing pleiotropic genes through integrative analyses of multiomics data, estimating how much of the phenotypic correlation between schizophrenia and immune diseases was explained by the genetic correlations we observed, and providing novel evidence for potential sex-specific pleiotropy between schizophrenia and immune disease.
Results
Defining immune risk variants
We identified immune-mediated diseases with robust GWAS findings using ImmunoBase (http://www.immunobase.org; accessed 7 June 2015), an online resource providing curated GWAS data for immune-related human diseases. These included the following 19 diseases: alopecia areata (AA), ankylosing spondylitis (AS), autoimmune thyroid disease (ATD), celiac disease (CEL), Crohn’s disease (CRO), inflammatory bowel disease (IBD), juvenile idiopathic arthritis (JIA), multiple sclerosis (MS), narcolepsy (NAR), primary biliary cirrhosis (PBC), primary sclerosing cholangitis (PSC), psoriasis (PSO), rheumatoid arthritis (RA), Sjögren’s syndrome (SJO), systemic lupus erythematosus (SLE), systemic sclerosis (SSC), type 1 diabetes (T1D), ulcerative colitis (UC), and vitiligo (VIT). Notably, the majority of IBD risk variants were also risk variants for CRO and/or UC. For 14 of these immune diseases (see Table 1), we also obtained full GWAS or Immunochip summary statistics allowing us to conduct additional polygenic risk scoring (PRS) [30, 32] and cross-trait Linkage Disequilibrium Score regression (LDSC) analyses [16].
Given that human leukocyte antigen (HLA) alleles within the major histocompatibility complex (MHC) region (chromosome 6: 25-34 Mb) account for a significant proportion of heritability of immune and inflammatory disorders [33], we considered HLA and non-HLA risk variants separately in our analyses. Within the MHC region we considered only the most strongly associated HLA variant (including SNPs, imputed HLA amino acid sites, and classical alleles) for each disease based on univariate analysis in previously published studies (see Table 2), because multivariate conditional analyses reporting adjusted effect sizes of independent HLA variants were not available for all immune diseases. Outside of the MHC region, we considered all non-HLA variants curated in ImmunoBase for each of the 19 immune diseases.
The number of genome-wide significant non-HLA risk loci for each of the 19 immune diseases varied from three (NAR) to 144 (IBD). Several variants were associated with more than one immune disease. In total we identified 581 unique variants (563 non-HLA variants and 18 HLA variants) associated with any immune disease at genome-wide significance. We refer to these variants as immune risk variants.
Identifying pleiotropic variants implicated in both immune disease and schizophrenia
First, we evaluated whether there was any evidence of overall risk allele sharing between each of the 19 immune diseases and schizophrenia using a binomial sign test. To do this, we used previously published findings from a GWAS conducted by the Schizophrenia Working Group of the Psychiatric Genomics Consortium [1, 11]. This GWAS represented a meta-analysis of 52 cohorts, comprising a total of 35,476 cases and 46,839 controls, and the full dataset is referred to here as the PGC2 study. Overall, the direction of effect for the sets of non-HLA SNPs associated with each of the 19 immune diseases at genome-wide significance was not shared with schizophrenia more than expected by chance (all binomial sign test p>0.05, S1 Fig). Thus, we did not observe evidence of risk allele sharing between any immune disease and schizophrenia when using a stringent genome-wide significance threshold to define immune risk variants. We also evaluated the collective association of 261 LD-independent, non-HLA immune risk variants associated with at least one of the 19 immune-mediated diseases, for which linkage disequilibrium (LD) Score and minor allele frequency (MAF) information were available in the European LD Score database [16]. We found significant deviation from the theoretical null in schizophrenia for immune risk SNPs (λ=1.46). However, when we compared the association of immune risk SNPs to that of similar randomly selected SNP sets (Supplementary Methods) we observed no evidence of enrichment (S2 Fig, p=0.66), indicating that immune risk SNPs were not associated with schizophrenia more than expected by chance given the polygenic nature of schizophrenia.
Next, we identified potential pleiotropic variants by evaluating the association of individual immune risk variants with schizophrenia. We considered SNPs associated with schizophrenia at p<8.6×10−5 (Bonferroni correction for 581 tests, 563 non-HLA and 18 HLA variants) to have pleiotropic effects. Given the size of the schizophrenia GWAS, we had over 80% power to detect pleiotropic SNPs assuming an OR≥1.12 in schizophrenia.
Within the MHC region, we observed four HLA risk alleles associated with both immune disease and schizophrenia, particularly in the class II HLA region (Table 2, S3 Fig). These HLA risk alleles were the strongest MHC region associations for AA (HLA-DRB1 #37 Asn), CEL (HLA-DQB1 #74 Ala), PSC (HLA-B*08:01), and SJO (HLA-DQB1*02:01). The presence of HLA-DRB1 #37 Asn conferred a protective association in both AA and schizophrenia, but the remaining HLA variants showed the opposite direction of effect in schizophrenia compared to immune disease (Table 2, S3 Fig). Notably, none of these four HLA variants were significantly associated with schizophrenia in previous conditional analyses [9, 11], suggesting that their association with schizophrenia may be driven by LD with other causal variants in the region rather than true pleiotropy. Thus, we did not focus additional analyses on these variants.
Outside of the MHC region, five immune risk variants showed potential pleiotropic effects, with the risk allele for immune disease also conferring risk for schizophrenia. These variants have been previously implicated in CRO (rs6738825, rs13126505, rs1734907 [34, 35]), MS (rs7132277 [36]), and CEL (rs296547 [37]). To evaluate the pleiotropic potential of these non-HLA variants, we used conditional and joint analysis (COJO) [38] to perform association analyses in the PGC2 schizophrenia GWAS conditioning on each of the five immune risk variants (S4 Fig). In the setting of true pleiotropy, no significant associations should remain after conditioning on the immune risk variants (statistically, all p>8.6×10−5). Consistent with pleiotropy, we observed no remaining associations with schizophrenia after conditioning on rs296547 (top SNP after conditioning: rs111530734, p=1.19×10−3), rs1734907 (top SNP after conditioning: rs11768688, p=9.79×10−4), and rs13126505 (top SNP after conditioning: rs112786981, p=4.58×10−4). Significant associations with schizophrenia remained after conditioning on rs6738825 (top SNP after conditioning: rs111744017, p=8.03×10−6) and rs7132277 (top SNP after conditioning: rs74240770, p=1.37×10−8), suggesting there were independent causal variants driving the associations in these regions for schizophrenia and immune disorders.
In order to prioritize genes underlying the identified pleiotropic SNPs (rs296547, rs1734907, rs13126505), we performed an integrative analysis of GWAS summary statistics with methylation quantitative trait loci (mQTL) and expression quantitative trait loci (eQTL) studies using SMR and HEIDI [39, 40] (Materials and Methods). Notably, rs296547 was not genotyped in the eQTL dataset, and we used rs404339 as a proxy SNP (r2=0.85 in 1000 Genomes Phase 3 CEU Population [41]) in SMR analyses of gene expression analyses for rs296547. We observed that rs1734907 was an mQTL (β = 0.47, P = 2.13×10−26) and eQTL (β = −0.24, P = 3.54×10−10) for EPHB4 in peripheral blood (S2 Table, Fig 1). Furthermore, we observed consistent pleiotropic associations for rs1734907 with schizophrenia and EPHB4 DNAm (βSMR = −0.14, PSMR = 3.58×10−5, PHEIDI = 0.12), schizophrenia and EPHB4 expression (βSMR = −0.28, PSMR = 2.63×10−4, PHEIDI = 0.17), and EPHB4 DNAm and EPHB4 expression (βSMR = 1.98, PSMR = 6.56×10−8, PHEIDI = 0.011). Thus, there was consistent association across molecular phenotypes and schizophrenia at the EPHB4 locus, suggesting this gene may be driving the association of rs1734907 in schizophrenia (Fig 1). Notably,TRIP6 is also a candidate functional gene underlying the association of rs1734907 with schizophrenia. We observed pleiotropic association for rs1734907 with schizophrenia and TRIP6 DNAm with inconsistent direction of effect (βSMR = 0.15, PSMR = 5.00×10−5, PHEIDI = 0.17 for probe cg18683606; βSMR = - 0.12, PSMR = 2.32×10−5, PHEIDI = 0.18 for probe cg27396824), a trend for association with schizophrenia and TRIP6 expression (βSMR = −0.33, PSMR = 6.38×10−4, PHEIDI = 0.14), but no significant association with TRIP6 DNAm and TRIP6 expression. The other pleiotropic SNPs (rs296547, rs13126505) did not demonstrate consistent localization to a particular gene across traits and molecular phenotypes (Table 3, S2 Table). We observed that rs296547 was an mQTL for C1orf106 (β = −1.04, P < 10−30), and found pleiotropic associations for rs296547 with schizophrenia and C1orf106 DNAm but no other phenotypes (Table 3, S2 Table). Similarly, we observed that rs13126505 was an mQTL (β = 0.49, P = 4.03×10−16) and eQTL (β = −0.27, P = 3.54×10−10) for SLC39A8, and found pleiotropic associations for rs13126505 with schizophrenia and SLC39A8 DNAm along with schizophrenia and SLC39A8 expression, but not SLC39A8 DNAm and expression (Table 3, S2 Table).
Detecting genetic correlations between immune disease and schizophrenia
Our immune risk variant set captured only those variants associated with immune diseases at genome-wide significance in current GWASs. Given the polygenicity of immune-related diseases, there are 100s to 1,000s of additional variants associated with each disease which have not yet been identified [42]. To evaluate sharing of risk alleles between immune diseases and schizophrenia using a broader set of variants, we used PRS [30, 32] and LDSC [16].
For each of the 14 immune diseases with available genome-wide summary statistics, we constructed genetic risk scores (GRSs) at a range of p-value thresholds (pT) as in previous studies [12], and tested for the association of these GRSs with schizophrenia in a refined subset of the PGC2 study (17,000 cases and 20,655 controls) which excluded samples shared with the immune disease GWASs. To benchmark our findings in immune diseases, we also analyzed human height [43] and included previously published PRS results for bipolar disorder [12]. We considered immune diseases with PRS p<0.002 at any pT to show significant genetic overlap with schizophrenia (Bonferroni correction for 14 immune diseases tested in both sexes, 0.05/(14*2) ≈ 0.002). Commonly used goodness-of-fit estimates obtained from PRS (such as βGRS and Nagelkerke’s pseudo-R2) lack meaningful interpretation, which makes it difficult to compare these estimates across studies [44]. For these reasons we chose to interpret the direction of effect (i.e. positive or negative correlation) obtained from βGRS, but not to interpret or compare the degree of genetic sharing between immune diseases and schizophrenia. For further details of our PRS approach, see Materials and Methods. Using PRS, we had over 80% power to detect genetic covariance with schizophrenia ranging from 0.02 to 0.03 for most of the immune diseases, although some showed less than 80% power in this range (PSO, SLE, VIT; S5 Fig).
As previously described, bipolar disorder PRSs were significantly associated with schizophrenia (p<1×10−50 at pT<1) [12]. Surprisingly, human height PRSs were also significantly associated with schizophrenia (p=1×10−11 at pT<1, S3 Table). Height was analyzed as a negative control based on its previously reported lack of genetic correlation with schizophrenia using LDSC [16]. Using PRS, we observed that genetic liability for increased height protected against schizophrenia (βGRS=-0.11 at pT<1). The significant inverse association of height PRSs with schizophrenia case-status we observed may reflect the greater sensitivity of this approach to subtle population stratification, sample sharing, and/or true genetic overlap.
Genetic scores including the HLA region were significant for CEL, NAR, PBC, PSO, RA, SLE, SSC, T1D, and UC (p<0.002 at multiple pT, S4 Table). Height was not included in these analyses, given that HLA variants have not been associated with height in previous GWAS [43]. With the exception of CEL (βGRS ≈ 0.04 at pT<5×10−8, 1×10−4, and 1×10−3), all immune diseases exhibited a positively associated PRS with schizophrenia case-status (all βGRS>0, S4 Table). For CEL, RA, SLE, and SSC only those PRSs constructed using the most stringent p-value cutoffs (5×10−8, 1×10−4, 1×10−3) were significantly associated with schizophrenia. To evaluate whether the HLA region alone was driving the observed genetic sharing, we constructed PRSs excluding this region. After excluding HLA variants, genetic scores for NAR, PBC, PSO, SLE, T1D, and UC remained significantly associated with schizophrenia (Table 4, S6 Fig). Because the genetic overlap between these six immune diseases and schizophrenia was not driven by a single HLA variant of large effect, we focused on these findings for the remainder of our analyses.
Given the potential sensitivity of PRS to artificial genetic overlap highlighted in our analysis of height, we wanted to assess whether cryptic sample sharing between the immune and schizophrenia GWASs could be driving the shared genetic liability that we observed. To do this, we conducted leave-half-out analyses. If the observed genetic overlap was driven by samples shared between certain schizophrenia cohorts and the immune disease GWASs, the GRS association should not be consistently observed across subsamples leaving out half of the schizophrenia cohorts. Across 1,000 subsamples (Ncases ranging from 3,985-13,074) leaving out a randomly selected 14 cohorts, we observed a high proportion of subsamples with GRSs significantly associated with schizophrenia (p<0.05 at pT<1) for height (0.99), NAR (0.72), PBC (0.95), PSO (0.84), SLE (0.97), T1D (0.95), and UC (0.70) suggesting our findings were not driven by sample sharing.
To further validate our finding of genetic overlap between schizophrenia and these six immune-mediated diseases using PRS, we applied an independent method (LDSC) for estimating genome-wide genetic correlation between traits that is robust to sample sharing [16]. For LDSC analyses, we used summary statistics from the 49 European-ancestry cohorts in the PGC2 study (31,335 cases and 38,765 controls) [1]. Unlike PRS, LDSC provides an interpretable and comparable estimation of genetic sharing between two traits in the form of genetic correlation (rg) values. Notably, LDSC is less sensitive than PRS and is not robust when applied to genetic data obtained from specialty chips (e.g. Immunochip) [16]. We did not carry T1D forward for LDSC analysis, due to failure of this dataset on quality control measures (liability scale h2 >1, likely secondary to inflated effective sample size due to genotyping on Immunochip). Given that this was a secondary analysis, we considered immune diseases with rg p<0.05 to show significant genetic overlap with schizophrenia.
As previously reported [16], our positive control (bipolar disorder) showed significant genetic overlap with schizophrenia (rg=0.75±0.05, p=8.5×10−60; Fig 2, Table 4). In contrast to our PRS results, but in agreement with previous findings [16], our negative control (height) showed no such overlap using LDSC (rg=-0.004±0.02, p=0.84; Fig 2, Table 4). With respect to immune diseases, LDSC confirmed significant genetic overlap with schizophrenia for PBC, PSO, SLE, and UC (rg=0.10-0.18, Fig 2, Table 4) indicating the association of GRSs for these diseases was not driven by shared samples. Notably, genetic correlations for PSO and SLE did not survive correction for the 14 tests performed (Table 4). We also observed significant genetic overlap with schizophrenia for NAR using LDSC, with the caveat that this dataset was genotyped using Immunochip and did not survive multiple testing correction (Fig 2, Table 4). Overall, LDSC provided consistent results for the immune diseases showing significant genetic sharing with schizophrenia by PRS.
Benchmarking genetic correlations between immune disease and schizophrenia with epidemiological data
To determine how much of the phenotypic correlation between schizophrenia and immune-mediated diseases was explained by the genetic correlations we observed, we benchmarked significant genetic correlations between schizophrenia and immune-mediated disorders relative to the expected phenotypic correlations from epidemiological data (Materials and Methods). Using incidence of immune diseases in schizophrenia reported in a large population-based study [3], we estimated phenotypic correlations between schizophrenia and PBC, PSO, SLE, and UC. We were unable to estimate phenotypic correlation for NAR and schizophrenia, given that there were no estimates in the literature of the incidence of NAR in schizophrenia. For PBC, PSO, and SLE we observed small positive genetic correlations with schizophrenia that were consistent with the epidemiological data (PBC: rg = 0.131 ± 0.05, rp = 0.112; PSO: rg = 0.182 ± 0.07, rp = 0.130; SLE: rg = 0.130 ± 0.05, rp = 0.048). For UC we observed a small positive estimate of genetic correlation (rg = 0.106 ± 0.04) while there was no strong evidence for any correlation between UC and schizophrenia in the epidemiological data (rp = −0.001). Importantly, while the MHC region contains risk factors for both schizophrenia and immune-mediated diseases, our genetic correlation estimates were obtained considering only SNPs outside of the MHC region due to unusual LD in this region [45].
Exploring sex-dependent genetic correlations between immune disease and schizophrenia
Given the significant sex bias of autoimmune diseases, with women at greater risk overall [46], we hypothesized that there may be sex-dependent genetic overlap between schizophrenia and some immune-mediated diseases. We therefore performed sex-stratified PRS, testing the association of height and immune disease GRSs with schizophrenia separately in males and females of the PGC2 study. Genetic scores for height showed significant association with schizophrenia in both males and females. Three of the immune diseases (PBC, PSO, T1D) with significant main effects showed sex-dependent effects, with greater signal among males (S5 Table). Additionally, although genetic scores for MS were not significantly associated with schizophrenia in the total sample there was significant association among males (R2=0.03, p=1.26×10−3 at pT<1; S5 Table).
Given the greater statistical power for the male subset of the schizophrenia GWAS, we performed simulations by selecting random subsamples of male cases and controls equal in size to the female sample (5,321 cases and 9,094 controls). If the stronger genetic overlap between schizophrenia and MS, PBC, PSO, and T1D among males was driven by the larger sample size rather than a true sex-dependent effect, there should be no consistent association of GRSs with schizophrenia in these subsamples. Across 1,000 subsamples, the proportion with significant GRSs (p<0.002 at pT<1) was high for PBC (0.94) and T1D (0.87), suggesting our finding of a greater pleiotropic effect among males for these diseases was not driven solely by lower statistical power among females; this was not the case for PSO (0.59) or MS (0.21).
Next, we performed formal statistical tests for an interaction between sex and genetic scores for these four immune diseases. We observed a nominally significant interaction for MS (p<0.05 at several pT; S5 Table), noting that this finding did not survive correction for multiple testing. The remaining immune diseases did not show significant sex interactions, although the direction of effect was consistent with a greater pleiotropic effect in males (S5 Table).
Discussion
Using a variety of statistical approaches, we provide evidence of shared genetic risk for schizophrenia and several immune diseases. Within the MHC region, we identified four HLA variants showing statistically significant association with schizophrenia. An important caveat is that these four variants were not the top variants in their respective regions of association with schizophrenia, and were not primary drivers of the MHC association in schizophrenia in stepwise conditional analyses [9]. Therefore, the biological significance of these particular HLA variants in schizophrenia is likely limited.
Outside of the MHC region, we identified three SNPs with pleiotropic effects - influencing risk for both celiac disease (CEL) (rs296547) or Crohn’s disease (CRO) (rs1734907, rs13126505) and schizophrenia. Integration of GWAS, mQTL, and eQTL data implicated C1orf106, SLC39A8, and EPHB4 or TRIP6 as functional candidates driving the pleiotropic association of rs296547, rs13126505, and rs1734907, respectively. Overall, our findings provide the strongest evidence for a model in which genetic variation at rs1734907 (∼85kb upstream of EPHB4) increases DNA methylation, upregulates EPHB4 expression, and decreases the risk of schizophrenia. While DNA methylation is classically associated with gene silencing, the effect of methylation on transcription depends on the genomic context [47]; for instance, methylation of silencers or insulators eliminates transcription-blocking activity thereby promoting gene expression [48, 49]. EPHB4 is a transmembrane tyrosine kinase receptor that coordinates cell movement via bidirectional intercellular signaling at sites of direct cell-to-cell contact [50]. In the brain, ephrin signaling mediates synaptic plasticity by initiating and stabilizing neuronal synapse formation (reviewed by [51]). An analogous role has not yet been discovered in the immune system, possibly due to the much shorter lifespan of immunological synapses between lymphocytes and antigen presenting cells (minutes) as compared to neuronal synapses (years) [52, 53]. Interestingly, ephrin signaling attenuates the migration responses of both neurons and immune cells toward chemoattractants in vitro [54, 55]. Thus, disrupted pathfinding may be a shared risk mechanism by which EPHB4 contributes to immune disease and schizophrenia. The hypotheses raised by our findings require further validation. If the association of rs1734907 with CRO and schizophrenia is robustly replicated in future GWASs, functional studies will be needed to investigate both the genetic mechanism by which rs1734907 (or a causal variant in LD with this SNP) influences EPHB4 transcription, and the biological mechanism by which increased EPHB4 expression influences susceptibility to CRO and schizophrenia. With the multi-kinase inhibitor dasatinib already on the market for treatment of chronic myeloid leukemia [56] and other EphB4 inhibitors currently in Phase II trials [57–60], the potential for future drug repurposing makes EPHB4 an attractive candidate for further investigation.
We observed genome-wide sharing of risk variants for schizophrenia and six immune diseases (narcolepsy (NAR), primary biliary cirrhosis (PBC), psoriasis (PSO), systemic lupus erythematosus (SLE), type 1 diabetes (T1D), and ulcerative colitis (UC)) using PRS, all of which have been previously reported to co-occur with schizophrenia in epidemiological studies [3, 5, 61]. The strongest evidence of shared genetic risk emerged for PBC, PSO, SLE, and UC, which also showed robust genetic correlation with schizophrenia using LDSC. With the exception of UC, the small positive genetic correlations observed between these immune diseases and schizophrenia (rg ∼ 0.1) were consistent with phenotypic correlations observed in epidemiological data. Thus, currently available genetic data suggest that shared genetic risk contributes to the co-occurrence of PBC, PSO, and SLE in schizophrenia. Possible explanations for this sharing of genetic risk include the presence of a subgroup of “autoimmune-like” schizophrenia cases and/or sharing of specific biological pathways between schizophrenia and these particular immune diseases.
To our knowledge, this is the first time that sex-dependent genetic correlation with immune diseases has been investigated in schizophrenia. We found nominal evidence of male-specific genetic correlation for multiple sclerosis (MS), and a stronger pleiotropic effect among males for PBC, PSO, and T1D although the latter were not statistically significant. Interestingly, animal studies indicate that sex hormones have opposing effects on predisposition to schizophrenia and autoimmunity; estrogen has been reported to protect against the development of schizophrenia [62], while androgens appear to protect against the development autoimmune diseases [63, 64]. We emphasize that our sex-dependent findings require validation in independent samples. If replicated, one possibility is that sex hormones modulate pathogenesis among genetically vulnerable individuals, making males more likely to develop schizophrenia and females more likely to develop autoimmune diseases.
Our work was subject to several important limitations. Firstly, genome-wide summary statistics were not available for all of the immune diseases, resulting in a more limited analysis of 14 diseases. For five of these diseases (CEL, juvenile idiopathic arthritis (JIA), MS, NAR, T1D) summary statistics were obtained from Immunochip rather than GWAS, providing incomplete coverage of the genome for comparison with schizophrenia and biasing the genetic correlation estimates obtained by LDSC. Secondly, GRSs for human height – analyzed as a negative control – showed stronger association with schizophrenia than any of the immune diseases. An inverse epidemiological relationship between height and schizophrenia has been reported [65, 66], consistent with our PRS findings. The reasons for the discrepancy between PRS and LDSC, which showed no genetic correlation between height and schizophrenia (as previously reported [16]) are unclear. One explanation is that PRS, which uses individual-level genotype data as opposed to summary statistics, is a more sensitive method to detect true genome-wide sharing of risk alleles. If this is the case, it raises a broader question regarding how much genetic overlap is expected across complex traits in general using the PRS approach. Recent work suggests that pleiotropy is pervasive across human diseases, and that this phenomenon is driven at least in part by the polygenic nature of complex traits [21]. If this is the case, the extreme polygenicity of human height (more than 100,000 common variants estimated to exert independent causal effects [67]) may be driving the pleiotropy we observed between height and schizophrenia using PRS. An alternative explanation that must be considered is that PRS may be more vulnerable to confounding by cryptic population stratification, LD, or sample sharing.
Despite these limitations, our work adds to a growing body of evidence suggesting that schizophrenia and immune diseases share genetic risk factors. There are conflicting reports in the literature with respect to the specific immune diseases demonstrating genetic overlap with schizophrenia, and the direction of effect (positive or negative genetic correlation). Genetic overlap with schizophrenia has been previously investigated for nine of the 19 immune diseases studied here. Genome-wide genetic correlation with schizophrenia has been previously reported for CRO [23–25, 27], MS [28], PBC [25], PSO [25, 29], rheumatoid arthritis (RA, both positive [23, 24] and negative [31] genetic correlations), SLE [24, 25], T1D [23], and UC [24–27] (see S1 Table for a summary of previous studies). Our results are consistent with previously reported genetic overlap between schizophrenia and PBC [25], PSO [25], SLE [24, 25], T1D [23], and UC [24, 25]. While we did not observe genetic correlation between schizophrenia and MS in the total sample, there was a significant sex-dependent effect with genetic correlation observed among males. We provide new evidence of genetic correlation with NAR (not previously investigated). Notably, we did not find any significant genetic correlation between schizophrenia and RA. Despite the robust inverse epidemiological association between schizophrenia and RA [8], the genetic association is less consistent. Using methods based on summary statistics (including PRS and LDSC), four previous studies reported no evidence of pleiotropy between schizophrenia and RA [8, 16, 25, 30], while two studies reported positive genetic correlation [23, 24]. Notably, Lee et al. reported an inverse genetic correlation – in keeping with the observed epidemiological effect – using restricted maximum likelihood (GREML), a method utilizing full genotype data which has greater statistical power to detect small pleiotropic effects than PRS or LDSC [31]. Given the modest and potentially sex-dependent genetic correlations observed in the present study, subtle differences in statistical power across studies using different statistical methods and GWAS datasets may explain these discrepant findings. As genetic samples continue to grow, and our understanding of the degree of genetic overlap expected among complex traits evolves, it will be worthwhile to revisit these analyses.
Overall, our analyses provide statistical evidence supporting extensive pleiotropy between immune diseases and schizophrenia. Our results highlight EPHB4, a transmembrane receptor that coordinates cell migration and has dual roles in immune cell and neuronal pathfinding, as a promising candidate for future functional studies. More broadly, our findings indicate that common genetic variants influencing the risk of immune diseases – in particular NAR, PBC, PSO, SLE, and UC – are also involved in schizophrenia. Studies identifying the cell types and biological pathways that may be driving this genetic overlap are needed, and will hopefully provide further insights into the pathophysiology of schizophrenia. In the meantime, our work supports the emerging hypothesis that pathogenic mechanisms are shared across immune and central nervous system disorders.
Materials and Methods
Samples and quality control
We used either imputed genotype data or summary statistics generated as described in the original GWASs. For sample details, see Table 1.
Schizophrenia dataset
We used data from the PGC2 study [1]. For analyses of non-HLA genome-wide significant risk variants for immune diseases we used publicly available summary statistics from the total dataset (52 cohorts; 35,476 cases and 46,839 controls) [1]. For PRS analyses we used all 36 European ancestry case-control cohorts with available individual-level genotype data (25,629 cases and 30,976 controls). For analyses including HLA variants we used a further refined 31 European ancestry case-control cohorts (20,253 cases and 25,011 controls) with high-quality coverage of the MHC region, as previously described [11].
Immune disease datasets
To estimate the extent of genetic overlap between schizophrenia and immune diseases, we obtained full GWAS or Immunochip summary statistics for 14 of the 19 immune diseases (five immune diseases were not included in PRS analyses due to lack of available summary statistics). We obtained publicly available summary statistics for ten immune diseases (see URLs): CEL [68], CRO [69], IBD [69], JIA [70], MS [36], NAR [71], RA [72], SLE [73], T1D [74], and UC [69]. For the following four immune diseases, we obtained summary statistics with permission from the authors: PBC [75], PSO [76], SSC [77], and VIT [78].
Testing the association of genome-wide significant risk alleles for 19 immune diseases in schizophrenia
For each of the 19 immune diseases, we defined risk loci outside of the MHC region (chromosome 6: 25-34 Mb) using curated GWAS results from ImmunoBase (http://www.immunobase.org; accessed 7 June 2015. For details, see Supplementary Methods). Notably, the majority of IBD risk variants were also risk variants for CRO and/or UC. Within the MHC region we considered only the most strongly associated HLA variant (including SNPs, imputed HLA amino acid sites, and classical alleles) for each disease based on univariate analysis in previously published studies (see Table 2), because multivariate conditional analyses reporting adjusted effect sizes of independent HLA variants were not available for all immune diseases. In total there were 581 unique variants (563 non-HLA variants and 18 HLA variants) associated with any immune disease at genome-wide significance.
First, we tested for shared direction of effect with schizophrenia among SNPs associated with each of the 19 immune diseases using the binomial sign test. Because some immune risk SNPs were associated with multiple diseases with inconsistent direction of effect, we could not evaluate shared direction of effect among the collective set of immune risk SNPs in schizophrenia.
Next, we evaluated the collective association of SNPs associated with any immune disease. First we extracted the p-values for a pruned set of 261 LD-independent, non-HLA immune risk SNPs with linkage disequilibrium (LD) Score and minor allele frequency (MAF) information were available in the European LD Score database [16] from the schizophrenia PGC2 GWAS. We then quantified enrichment of these immune risk SNP associations in schizophrenia using the genomic inflation value λ. We obtained an empirical enrichment p-value by comparing this to λ values from 1,000 equal-sized sets of SNPs drawn from the schizophrenia GWAS summary data, and matched to the immune SNP set for MAF and LD score as these parameters are correlated with GWAS test statistics (see Supplementary Methods for details).
Finally, we evaluated the association of each of the 581 variants with schizophrenia using previously published association results for non-HLA [1] and HLA variants [11]. We considered SNPs associated with schizophrenia at p<8.6×10−5 (Bonferroni correction for 581 tests, 563 non-HLA and 18 HLA variants) to have pleiotropic effects.
To evaluate the pleiotropic potential of immune risk variants significantly associated with schizophrenia, we performed conditional and joint analysis (COJO) using GCTA [79]. Specifically, we used COJO to perform association analyses in the PGC2 schizophrenia GWAS conditioning on the immune risk variants of interest (i.e. SNPs that were significantly associated with both an immune disease and schizophrenia). In the setting of true pleiotropy, no significant associations with schizophrenia should remain after conditioning on these immune risk variants (statistically, all p>8.6×10−5). We used the 1000 Genomes Phase 3 European dataset as a reference panel to calculate LD between SNPs.
To prioritize genes and regulatory elements driving the pleiotropic GWAS loci we identified (associated with both immune disease and schizophrenia, see Table 3), we followed the analytic approach described by Wu et al. [40]. This approach integrates summary statistics from independent -omics methylation quantitative trait loci (mQTL) studies, expression quantitative trait loci (eQTL) studies, and GWAS to identify SNPs associated with gene expression, DNA methylation, and disease through shared genetic effects.
We obtained mQTL and eQTL data used in Wu et al. [40] for genetic regions within a 2Mb window of each pleiotropic SNP. These data and the quality control measures applied have been described in detail elsewhere [40]. Briefly, mQTL summary-level SNP data were from a meta-analysis of the Brisbane Systems Genetics Study [80] and Lothian Birth Cohorts of 1921 and 1936 [81], which comprised 1,980 individuals with DNA methylation measured in peripheral blood. eQTL summary-level SNP data were from the Consortium for the Architecture of Gene Expression (CAGE) study [82], which comprised 2,765 individuals with gene expression levels measured in peripheral blood. GWAS summary-level SNP data for schizophrenia was from the PGC2 study [1].
We applied summary data-based Mendelian randomization (SMR) using GCTA [79] to test for shared associations between the pleiotropic SNPs with DNAm probes and gene expression probes, DNAm probes and schizophrenia, and gene expression probes and schizophrenia. We included DNAm and gene expression probes within 2Mb of the pleiotropic SNPs. We considered significant associations as those with PSMR < 1.30×10−4 (0.05/385 tagged genes) for mQTLs and PSMR < 4.31×10−4 for eQTLs (0.05/116 tagged genes). Next, we applied the heterogeneity in dependent instruments (HEIDI) test [39] using GCTA [79] to evaluate whether significant shared associations between DNAm, gene expression and schizophrenia were driven by linkage (i.e. separate causal variants in LD exerting genetic effects on DNAm, gene expression, and schizophrenia) or a shared pleiotropic causal variant. We considered genetic effects that passed the HEIDI test (PHEIDI > 0.01) to be driven by a single causal variant. We looked for consistent SMR and HEIDI results across GWAS, mQTL, and eQTL studies to prioritize genes for future functional studies.
Testing the association of polygenic risk scores for 14 immune diseases in schizophrenia
To evaluate whether common variants influencing risk of immune diseases collectively contribute to schizophrenia, we used PRS [30, 32]. To benchmark the amount of genetic overlap between schizophrenia and immune disease, we included previously published results for bipolar disorder as a positive control [12]. We used human height [43] as a negative control because – despite the inverse epidemiological relationship between height and schizophrenia previously reported [65, 66] – a prior study using cross-trait LDSC reported no genetic correlation with schizophrenia [16].
For 14 immune diseases with available genome-wide summary statistics we performed PRS at a range of p-value thresholds (pT) as in previous studies [12]: 5×10−8, 1×10−4, 1×10−3, 0.01, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, and 1.0 (which included all LD-independent SNPs, Table 1). Due to extensive LD in the HLA region, we performed analyses both including the top HLA variant and excluding the HLA region. At each pT, we constructed GRSs for each individual i in the schizophrenia cohort for each immune disease h by calculating the sum of risk-allele dosages (g) weighted by their effect sizes (β) for that immune disease: where M iterates over all known risk alleles for disease h, βM,h is the effect size (log odds ratio) of M in disease h, and gM,i is the risk-allele dosage of M in individual i. We then performed logistic regression in R [83] using the stats package [83] to evaluate the association between schizophrenia case-status and GRSs for each immune disease. As in previous studies, statistical significance of the GRSs was estimated based on their logistic regression coefficient [12, 30]. Variance in schizophrenia case-status explained by the GRSs was estimated using the deviation in liability-scale R2 between a null model (including 10 ancestry-informative principal components and study site) and the full model (including GRSs in addition to these covariates), calculated as previously described [44] assuming a population prevalence of schizophrenia of 1%. We also estimated Nagelkerke’s pseudo-R2 using the fmsb package [84]. We considered immune diseases with GRS p<0.002 at any pT to show significant genetic overlap with schizophrenia (Bonferroni correction for 14 immune diseases tested in both sexes, 0.05/(14*2)=0.002). As in previous studies [12, 30] we did not use Bonferroni correction for the number of p-value thresholds, as these tests are highly correlated.
We excluded eight schizophrenia cohorts using Wellcome Trust Case Control Consortium (WTCCC) controls, due to the use of these samples in the immune disease GWASs. The total schizophrenia sample analyzed by PRS included 37,655 subjects (28 cohorts; 17,000 cases and 20,655 controls). Sex-stratified and formal sex-PRS interaction analyses were performed among the subset of subjects with known sex (9,787 male cases and 9,284 male controls; 5,231 female cases and 9,094 female controls). For details of PRS, see Supplementary Methods and Table 1.
Estimating the degree of genetic correlation between schizophrenia and 14 immune diseases
To validate our PRS results and obtain genetic correlation (rg) estimates, we performed a secondary analysis using cross-trait LDSC for immune-mediated diseases with significant PRS associations with schizophrenia [16]. Cross-trait LDSC estimates the genetic correlation between two traits using GWAS summary statistics. Similar to the PRS analyses described above, we benchmarked the genetic correlations observed for immune diseases by analyzing bipolar disorder [85] as a positive control and human height [43] as a negative control.
The statistical framework for cross-trait LDSC has been described in detail previously [16]. Briefly, LDSC leverages the relationship between LD and association test statistics to estimate heritability as the slope of the regression of z-scores against LD scores [86]. Cross-trait LDSC is a bivariate extension of this method which estimates genetic covariance as the slope of the regression of the products of z-scores against LD scores using the following equation [16]: where zij denotes the z score for study i and SNPj, 𝓁j is the LD score [86], Ni is the sample size for study i, ϱg is the genetic covariance, M is the number of SNPs in the reference panel with MAF between 5% and 50%, NS is the number of individuals included in both studies, and ϱ is the phenotypic correlation among the NS overlapping samples. Genetic covariance ϱg is estimated by regressing z1j z2 against , and multiplying the resulting slope by M. Statistical significance is assessed using block jackknifing over 200 equally sized blocks of SNPs [16]. Importantly, the MHC region is excluded from LDSC analyses due to its unusual LD structure and genetic architecture [45].
Because LDSC is robust to sample sharing across GWAS [16], we used summary statistics from the 49 European-ancestry cohorts in the PGC2 study (31,335 cases and 38,765 controls) [1]. We used LD Scores from the “eur_w_ld_chr/” files available from https://data.broadinstitute.org/alkesgroup/LDSCORE, computed using 1000 Genomes Project [87] Europeans as a reference panel as previously described [45]. To ensure we were using well-imputed SNPS we filtered all GWAS as previously described [16], including limiting the analysis to HapMap 3 [88] SNPs as implemented in the LDSC script munge_sumstats.py (https://github.com/bulik/ldsc). We estimated liability scale h2 for each trait using previously reported prevalence estimates (S6 Table), and removed datasets with h2>1. Given that this was a secondary analysis, we considered traits with rg p<0.05 to have significant genetic correlation with schizophrenia.
Benchmarking with epidemiological data
To determine how much of the phenotypic correlation between schizophrenia and immune-mediated diseases was explained by the genetic correlations we observed, we used the approach previously described by Lee et al. [31]. Briefly, we benchmarked our significant genetic correlation estimates between schizophrenia and NAR, PBC, PSO, SLE and UC relative to the expected phenotypic correlations from epidemiological data. We obtained estimates of the population risk of schizophrenia (KSCZ), the population risk of each immune disease (KIMMUNE), and the probability of each immune disease among patients with schizophrenia (KIMMUNE | SCZ) from the literature as referenced in S6 Table. We estimated the phenotypic correlation between schizophrenia and the immune disease of interest (RSCZ-IMMUNE) using the following formula, as derived by Lee et al. [31] assuming that the phenotypic liabilities of schizophrenia (lSCZ) and immune disease (lIMMUNE) follow a bivariate normal distribution with mean=0 and standard deviation=1: where:
tSCZ is the liability threshold for schizophrenia:
tIMMUNE is the liability threshold for immune disease:
tIMMUNE | SCZ is the liability threshold for immune disease in those with schizophrenia:
dSCZ is the “height” of the normal distribution at the schizophrenia liability threshold:
iSCZ is the mean phenotypic liability of those with schizophrenia:
Statistical power
Power to detect association of individual non-HLA and HLA immune risk variants in schizophrenia was calculated using the Genetic Power Calculator [89] assuming a risk allele frequency (RAF) of 0.05, disease prevalence of 1%, and significance threshold (α) of 8.6×10−5. Power for PRS was evaluated using AVENGEME [90, 91], assuming disease and genetic parameters detailed in S6 Table.
URLs
LD Score database: ftp://atguftp.mgh.harvard.edu/brendan/1k_eur_r2_hm3snps_se_weights.RDS
GWAS summary statistics:
CEL
CRO, IBD, UC
ftp://ftp.sanger.ac.uk/pub/consortia/ibdgenetics/iibdgc-trans-ancestry-filtered-summary-stats.tgz
JIA
MS
https://www.immunobase.org/downloads/protected_data/GWAS_Data/hg19_gwas_ms_im_sgc_4_19_1.tab.gz
NAR
RA
http://www.broadinstitute.org/ftp/pub/rheumatoid_arthritis/Stahl_etal_2010NG/
SLE
https://www.immunobase.org/downloads/protected_data/GWAS_Data/hg19_gwas_sle_be_ntham_4_20_0.tab.gz
T1D
Acknowledgements
We thank all study participants, and all research staff at the numerous study sites that made this research possible. Computations were performed on the Genetic Cluster Computer (GCC) hosted by SURFsara (www.surfsara.nl). We thank ImmunoBase for curating genetic risk factors for the immune diseases investigated. We thank the Canadian-US PBC Consortium, the Italian PBC Genetics Study Group and the UK-PBC Consortium for sharing their summary data for the PBC GWAS. We thank the International Inflammatory Bowel Disease Genetics Consortium (IIBDGC) for providing publicly available genome-wide summary statistics for CRO, UC, and IBD. We also thank the International Multiple Sclerosis Genetics Consortium (IMSGC) and the investigators leading the GWASs for CEL, JIA, RA, SLE, and T1D for providing genome-wide summary statistics from their studies on ImmunoBase.
This research was supported in part by a number of funding sources. This research uses resources provided by the Genetic Association Information Network (GAIN), obtained from the database of Genotypes and Phenotypes (dbGaP) found at http://www.ncbi.nlm.nih.gov/gap through dbGaP accession number phs000021.v3.p2; samples and associated phenotype data for this study were provided by the Molecular Genetics of Schizophrenia Collaboration (PI: Pablo V. Gejman, Evanston Northwestern Healthcare (ENH) and Northwestern University, Evanston, IL, USA). JG Pouget is supported by Fulbright Canada, the Weston Foundation, and Brain Canada through the Canada Brain Research Fund - a public-private partnership established by the Government of Canada. B Han is supported by the National Research Foundation of Korea (NRF) (grant 2016R1C1B2013126) and the Bio & Medical Technology Development Program of the NRF (grant 2017M3A9B6061852) funded by the Korean government, Ministry of Science and ICT (MSIT). HM Ollila is supported by the Finnish Cultural Foundation and Academy of Finland (grant 309643). J Martin is supported by grants SAF2015-66761-P from the Spanish Ministry of Economy and Competitiveness and P12-BIO-1395 from from Consejería de Innovación, Ciencia y Tecnología, Junta de Andalucía (Spain). Y Jin, SA Santorico, and R Spritz are supported by the U.S. National Institutes of Health (NIH) grants R01AR045584, R01AR056292, X01HG007484, and P30AR057212 and by institutional research funding IUT20-46 from the Estonian Ministry of Education and Research. MD Mayes is supported by the U.S. NIH grants N01AR02251 and R01AR05528. S Raychaudhuri is supported by the U.S. NIH grants 1R01AR063759, 1R01AR062886, 1UH2AR067677-01, and U19AI111224-01 and Doris Duke Charitable Foundation grant 2013097. J Knight is the Joanne Murphy Professor in Behavioural Science. Funding for the GAIN schizophrenia sample was provided by the U.S. NIH grants R01 MH67257, R01 MH59588, R01 MH59571, R01 MH59565, R01 MH59587, R01 MH60870, R01 MH59566, R01 MH59586, R01 MH61675, R01 MH60879, R01 MH81800, U01 MH46276, U01 MH46289 U01 MH46318, U01 MH79469, and U01 MH79470 and the genotyping of samples was provided through GAIN. The funding sources did not influence the study design, data analysis, or writing of this manuscript.
Footnotes
^ Membership of the Schizophrenia Working Group of the Psychiatric Genomics Consortium is provided in the Supporting Information.
References
- 1.↵
- 2.↵
- 3.↵
- 4.
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.
- 16.↵
- 17.
- 18.↵
- 19.↵
- 20.
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.
- 59.
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.
- 93.
- 94.
- 95.
- 96.
- 97.
- 98.
- 99.
- 100.
- 101.
- 102.
- 103.
- 104.
- 105.
- 106.