A molecule-based genetic association approach implicates a range of voltage-gated calcium channels associated with schizophrenia

Abstract

Traditional GWAS have successfully detected genetic variants associated with schizophrenia. However, only a small fraction of heritability can be explained. Gene-set/pathway based methods can overcome limitations arising from single SNP-based analysis, but most of them place constraints on size which may exclude highly specific and functional sets, like macromolecules. Voltage-gated calcium (Ca_v) channels, belonging to macromolecules, are composed of several subunits whose encoding genes are located far away or even on different chromosomes. We combined information about such molecules with GWAS data to investigate how functional channels associated with schizophrenia. We defined a biologically meaningful SNP-set based on channel structure and performed an association study by using a validated method: SNP-set (Sequence) Kernel Association Test. We identified 8 subtypes of Ca_v channels significantly associated with schizophrenia from a subsample of published data (N = 56,605), including the L-type channels (Ca_v1.1, Ca_v1.2, Ca_v1.3), P-/Q-type Ca_v2.1, N-type Ca_v2.2, R-type Ca_v2.3, T-type Ca_v3.1 and Ca_v3.3. Only genes from Ca_v1.2 and Ca_v3.3 have been implicated by the largest GWAS (N = 82,315). Each subtype of Ca_v channels showed relatively high chip heritability, proportional to the size of its constituent gene regions. The results suggest that abnormalities of Ca_v channels may play an important role in the pathophysiology of schizophrenia and these channels may represent appropriate drug targets for therapeutics. Analyzing subunit-encoding genes of a macromolecule in aggregate is a complementary way to identify more genetic variants of polygenic diseases. This study offers the potential of power for discovery the biological mechanisms of schizophrenia.

Introduction

Schizophrenia is a highly heritable complex disease (Lichtenstein et al. 2009). The biological underpinnings of schizophrenia remain an enigma, making prevention difficult and delaying development of better treatment alternatives (Van Os and Kapur 2009). Recently, advances in technology and the establishment of an international consortium, the Psychiatric Genomics Consortium (PGC), have made it possible to perform genome-wide association studies (GWAS) involving more than a hundred thousand individuals. The latest study from PGC has reported 108 independent genomic regions associated with schizophrenia (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). However, the variants identified can only explain a small fraction of the estimated heritability (Giusti-Rodríguez and Sullivan 2013; Goldstein 2009; Ripke et al. 2013; Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014), and the functional consequences of these variants remain largely uncharacterized. These problems may originate from inherent limitations of the GWAS methodology: The mass univariate testing approach requires an extremely stringent significance threshold to control false positives, thus reducing power; Genetic heterogeneity further complicate interpretation in large meta-analysis; Connecting SNP markers to the causal variants they represent is not straightforward; And, robust, efficient methods for detecting interactions among genetic variants remain elusive.

Gene-based, and gene-set/pathway based methods provide promising alternatives to overcome certain limitations of GWAS (Askland et al. 2012). Typically, genetic variants within or near to a gene are aggregated and tested for associations with a disease (Liu et al. 2010). Gene-set/pathway based analyses aggregate functionally related genes, providing a potentially powerful and biologically oriented bridge between genotypes and phenotypes (Ramanan et al. 2012; Wang et al. 2010). These methods, complementary to GWAS, have several advantages: They can reduce the number of tests performed; They may reduce the impact of genetic heterogeneity across cohorts; And they can facilitate the interpretation of findings. On the other hand, they also have limitations: Genes typically work in concert with one another (Liu et al. 2010), thus gene-based methods cannot take into account the joint effect among genes; The organization of pathways is typically derived from experiments of model organisms or predicted from mathematical models so uncertainties may be present (Bauer‐Mehren et al. 2009); The mechanism of the pathways is rarely clear (Khatri et al. 2012); And most published gene-set/pathway analyses place constraints on size from ten to a few hundred genes (Ramanan et al. 2012). Restriction to pathways with more than ten genes may exclude highly specific and potentially informative functional SNP sets, like macromolecules.

A macromolecule is a very large molecule created by polymerization of multiple smaller subunits. Voltage-gated calcium (Ca_v) channels that belong to macromolecules, are pore-forming membrane proteins involved in diverse physiological processes including depolarization of neuronal action potentials, neurotransmitter release, neuronal excitability and intracellular signaling(Simms and Zamponi 2014). Before interesting GWAS findings emerged, they have already received considerable physiological investigations in psychiatric and neurological disorders due to their importance to brain function (Catterall 2000; Simms and Zamponi 2014). Ca_v channels are key mediators of calcium entry into neurons (Turner et al. 2011) and calcium signaling is involved in major molecular hypothesis of schizophrenia such as dopamine, glutamatergic and GABAergic hypothesis (Lidow 2003). In fact, calcium signaling dysfunction has been suggested as a unifying pathological mechanism in schizophrenia (Lidow 2003). Thus, Ca_v channels gene variants are of large interest in relationship to schizophrenia and we chose to perform the macromolecular analysis of functional Ca_v channels.

Recent GWAS have identified several associated neuronal ion channel genes (e.g. CACNA1C, CACNB2, CACNA1I, KCNB1, HCN1, CHRNA3, CHRNA5, CHRNB4) (Cross-Disorder Group of the Psychiatric Genomics Consortium 2013; Ripke et al. 2013). In particular, associations at CACNA1C, CACNB2 and CACNA1I, which encode Ca_v channel subunits, extend previous findings implicating members of Ca_v channels in schizophrenia (Hamshere et al. 2013; Ripke et al. 2013). Ca_v channels can either be monomers (one subunit), or heteromultimers (three or four subunits). Although these subunits physically bind together to form a channel, their encoding genes are located in different regions of a chromosome or even on different chromosomes. For example, in the Ca_v1.1 channel (Bannister and Beam 2013), the α₁ subunit gene CACNA1S, α₂δ subunit gene CACNA2D1, β subunit gene CACNB1 and γ subunit gene CACNG1 are located at chromosomal bands 1q32, 7q21-q22, 17q21-q22 and 17q24, respectively (Fig. 1). Due to the limitations of gene-based and gene-set based analysis mentioned above, it is possible that taking the macromolecules (Ca_v channels) as a joint entity can explain more for the risk of schizophrenia than one single locus alone.

Fig. 1.

Molecular organization of voltage-gated calcium channels and chromosome locations of their subunit-coding genes.

Most Ca_v channels are multi-subunit structure (containing 3 or 4 subunits, α₁, β, α₂δ, with or without γ subunits), but T-type Ca_v channels only have the α₁ subunit. In one specific channel, the subunits are physically bound together, but their encoding genes are localized far apart or even on different chromosomes. Nine autosomal genes (CACNA1A, CACNA1B, CACNA1C, CACNA1D, CACNA1E, CACNA1G, CACNA1H, CACNA1I, CACNA1S) encode α₁ subunit (connected by red lines), four genes (CACNB1, CACNB2, CACNB3, CACNB4) encode β subunits (connected by blue lines), four genes (CACNA2D1, CACNA2D2, CACNA2D3, CACNA2D4) encode α₂δ subunit (connected by green lines) and eight genes (CACNG1, CACNG2, CACNG3, CACNG4, CACNG5, CACNG6, CACNG7, CACNG8) encode γ subunit (connected by grey lines). The numbers 1, 2, 3, 7, 9, 10, 12, 16, 17, 19 & 22 represent chromosome numbers.

We defined a SNP set from single channel genes and investigated how this biologically functional unit is associated with schizophrenia, using the accessible PGC schizophrenia GWAS data (N = 56,605: 25,629 cases and 30,976 controls) divided into a discovery and a replication sample. We applied the SNP-set (Sequence) Kernel Association Test (SKAT) (Wu et al. 2010) and identified significant associations in eight subtypes of Ca_v channels (Ca_v1.1, Ca_v1.2, Ca_v1.3, Ca_v2.1, Ca_v2.2, Ca_v2.3, Ca_v3.1 and Ca_v3.3). In contrast, only genes (CACNA1C, CACNB2 and CACNA1I) from two subtypes were implicated by the original GWAS despite its larger sample (N = 82,315). These findings show the potential of the macromolecule approach to identify the possible etiology of diseases, and suggest that abnormalities of Ca_v channels may play an important role in the pathophysiology of schizophrenia.

Materials and Methods

Ca_v Genes

A total of 26 genes encoding subunits of Ca_v channels can be classified into 4 groups (Table 1) according to the types of subunits they encode (Catterall 2000; Simms and Zamponi 2014). Genes CACNA1A, CACNA1B, CACNA1C, CACNA1D, CACNA1E, CACNA1F, CACNA1G, CACNA1H, CACNA1I, CACNA1S encode the α₁ subunits, CACNA2D1, CACNA2D2, CACNA2D3, CACNA2D4 encode the α₂δ subunits, CACNB1, CACNB2, CACNB3, CACNB4 encode the β subunits and CACNG1, CACNG2, CACNG3, CACNG4, CACNG5, CACNG6, CACNG7, CACNG8 encode the γ subunits. We only analyzed genes located on the autosomes, so the gene CACNA1F on the X-chromosome was excluded.

Table 1.

Gene-level test result from discovery and validation stages.

Gene Name	Type of Encoding Subunit	Stage1 BH_SKAT P	Stage2 BH_SKAT P	Combined dataset BH_SKATP
CACNA1A	α1A	4.66E-01	3.80E-01	2.29E-01
CACNA1B	α1B	7.15E-01	1.75E-01	1.75E-01
CACNA1C	α1C	2.24E-04*	2.42E-12*	3.07E-18*
CACNA1D	α1D	5.85E-01	7.63E-01	5.53E-01
CACNA1E	α1E	4.48E-01	7.49E-02	8.85E-03*
CACNA1G	α1G	8.94E-03*	1.75E-01	3.41E-03*
CACNA1H	α1H	5.60E-01	8.16E-01	3.29E-01
CACNA1I	α1I	3.75E-04*	2.32E-04*	9.88E-09*
CACNA1S	α1S	2.13E-01	2.64E-01	1.26E-01
CACNA2D1	α2δ1	5.85E-01	8.42E-02	1.26E-01
CACNA2D2	α2δ2	5.60E-01	7.49E-02	1.94E-01
CACNA2D3	α2δ3	4.48E-01	8.42E-02	8.01E-02
CACNA2D4	α2δ4	4.48E-01	1.60E-01	7.15E-02
CACNB1	β1	7.15E-01	1.75E-01	1.53E-01
CACNB2	β2	3.55E-02*	6.73E-02	2.41E-05*
CACNB3	β3	4.66E-01	1.92E-01	1.60E-01
CACNB4	β4	6.68E-01	1.86E-01	1.75E-01
CACNG1	γ1	6.48E-01	1.75E-01	1.53E-01
CACNG2	γ2	4.48E-01	1.75E-01	2.91E-01
CACNG3	γ3	4.66E-01	1.60E-01	1.26E-01
CACNG4	γ4	8.47E-01	7.49E-02	1.26E-01
CACNG5	γ5	4.66E-01	1.87E-01	1.26E-01
CACNG6	γ6	4.66E-01	1.75E-01	4.98E-01
CACNG7	γ7	5.75E-01	1.75E-01	1.26E-01
CACNG8	γ8	4.66E-01	2.65E-01	1.75E-01

^*:P-value < 0.05 after correction.

Stage1: discovery phase; Stage 2: validation phase; BH: Benjamini Hochberg; SKAT: SNP-set (Sequence) kernel association test.

Genotype data

Due to IRB restrictions from some sub-studies in PGC, we used the largest accessible PGC schizophrenia data which contains 36 case-control sub-studies (N = 56,605; 25,629 cases and 30,976 controls compared to 52 sub-studies and N=82,315 in the primary study) (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). Quality control and imputation were performed by the PGC Statistical Analysis Group for each dataset separately. Briefly, SNP meets with following conditions were retained: SNP missingness < 0.05, SNP Hardy-Weinberg equilibrium P > 1×10⁻⁶ in controls or P > 1×10⁻¹⁰ in cases. Samples with missing rate > 0.05 were removed. After quality control, the remaining genotypes were imputed using SHAPEIT2/IMPUTE2 (Delaneau et al. 2014; Howie et al. 2012) based on the full 1000 Genomes Project dataset (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). To evaluate the replicability of our analysis, we selected out the data used in the first phase of PGC (PGC1) as a discovery sample (10,616 cases and 10,315 controls), and used the rest as replication sample (15,013 cases and 20,661 controls). In addition, we also used combined samples from both discovery and replication stages. We first merged the best-guessed genotype data (imputation information score > 0.8 and minor allele frequency > 0.05) across 36 sub-studies, and then, performed the second round of quality controls using parameters SNP-missingness < 0.05 and minor allele frequency > 0.05. To control the impact of population stratification on our analysis, we computed the first 20 principal components based on the merged and quality controlled genotype data by using the program EigenSoft (Price et al. 2006). Since some Ca_v genes are close together in genomic position (for example, CACNG6, CACNG7 and CACNG8) it is possible that some SNPs may be assigned to more than one genes. In order to avoid the such undesired bias, we annotated SNPs to the closest gene (GENCODEv1.9) based on genomic positions that were derived from the human genome assembly build hg19. Then based on the SNPs list, the genotypes of the 25 Ca_v genes were extracted.

Ca_v channels can either be monomers (only the α₁ subunit), or heteromultimers (three subunits α₁, β, α₂δ; or four subunits α₁, β, α₂δ, γ). Great diversity of Ca_v channels allows them to fulfill highly specialized roles in specific neuronal subtypes (Simms and Zamponi 2014). Thus, for each α₁ subunit (principal subunit for classifying subtypes of Ca_v channels), co-assembly of a variety of ancillary subunits (β, α₂δ, γ) exists (Table 2). In some Ca_v channels, the ancillary subunit types are not completely known. So for channel-level association analysis, we test all of the possible combinations based on the current literatures (Buraei and Yang 2010; Catterall 1996; Davies et al. 2010; Hofmann et al. 2014; Schlick et al. 2010). According to different subunit gene combinations (3 or 4 genes per set), genotypes of the genes consisting of a Ca_v channel were concatenated. Therefore, each SNP set is corresponding to one functional channel that exists in nature.

Table 2.

Channel-level test results.

Channel Name	Subunits Combination	Stage1 BH_SKAT P	Stage2 BH_SKAT P	Combined datasets BH_SKAT P
Cav1.1	α1S β1 α2δ1 γ1	4.63E-01	2.78E-02*	3.54E-02*
Cav1.2	α1C β1 α2δ1	9.56E-04*	5.85E-12*	1.51E-16*
	α1C β1 α2δ2	5.09E-05*	8.42E-14*	1.62E-19*
	α1C β2 α2δ1	6.21E-05*	8.87E-13*	1.31E-20*
	α1C β2 α2δ1 γ1	6.21E-05*	8.05E-13*	1.16E-20*
	α1C β2 α2δ1 γ2	5.70E-05*	6.13E-13*	1.13E-20*
	α1C β2 α2δ1 γ3	6.21E-05*	4.66E-13*	5.56E-21*
	α1C β2 α2δ1 γ4	6.93E-05*	6.03E-13*	1.07E-20*
	α1C β2 α2δ1 γ5	6.21E-05*	6.93E-13*	6.59E-21*
	α1C β2 α2δ1 γ6	6.21E-05*	7.04E-13*	1.31E-20*
	α1C β2 α2δ1 γ7	6.21E-05*	8.05E-13*	1.13E-20*
	α1C β2 α2δ1 γ8	6.21E-05*	8.36E-13*	1.13E-20*
	α1C β2 α2δ2	6.66E-06*	8.72E-14*	1.75E-22*
	α1C β2 α2δ2 γ1	6.66E-06*	8.42E-14*	1.64E-22*
	α1C β2 α2δ2 γ2	6.66E-06*	8.42E-14*	1.61E-22*
	α1C β2 α2δ2 γ3	6.66E-06*	8.42E-14*	1.16E-22*
	α1C β2 α2δ2 γ4	7.22E-06*	8.42E-14*	1.61E-22*
	α1C β2 α2δ2 γ5	6.66E-06*	8.42E-14*	1.16E-22*
	α1C β2 α2δ2 γ6	6.66E-06*	8.42E-14*	1.64E-22*
	α1C β2 α2δ2 γ7	6.66E-06*	8.42E-14*	1.61E-22*
	α1C β2 α2δ2 γ8	6.66E-06*	8.42E-14*	1.61E-22*
	α1C β2 α2δ3	3.72E-05*	2.65E-12*	2.67E-20*
	α1C β2 α2δ4	6.66E-06*	1.49E-13*	1.16E-22*
	α1C β3 α2δ1	9.49E-04*	5.85E-12*	1.51E-16*
	α1C β3 α2δ2	5.09E-05*	8.42E-14*	1.62E-19*
	α1C β4 α2δ1	3.50E-03*	7.39E-11*	8.50E-15*
	α1C β4 α2δ2	7.72E-04*	4.67E-11*	6.26E-15*
Cav1.3	α1D β3 α2δ1	5.71E-01	6.08E-02	8.46E-02
	α1D β3 α2δ2	5.71E-01	1.85E-01	3.93E-01
	α1D β3 α2δ3	3.17E-01	5.79E-02	4.26E-02*
	α1D β3 α2δ4	4.39E-01	3.18E-01	1.21E-01
	α1D β4 α2δ1	6.41E-01	4.63E-02*	6.17E-02
	α1D β4 α2δ2	6.78E-01	1.22E-01	2.00E-01
	α1D β4 α2δ3	4.07E-01	4.36E-02*	3.54E-02*
	α1D β4 α2δ4	5.71E-01	1.72E-01	9.55E-02
Cav2.1	α1A β1 α2δ1	4.99E-01	4.00E-02*	4.42E-02*
	α1A β4 α2δ1	5.71E-01	3.28E-02*	3.72E-02*
	α1A β4 α2δ2	5.71E-01	7.90E-02	1.02E-01
	α1A β4 α2δ3	3.49E-01	3.27E-02*	2.13E-02*
	α1A β4 α2δ4	4.69E-01	1.13E-01	4.99E-02*
Cav2.2	α1B β1 α2δ1	6.41E-01	2.34E-02*	3.92E-02*
	α1B β1 α2δ2	7.14E-01	3.28E-02*	1.02E-01
	α1B β1 α2δ3	3.58E-01	2.34E-02*	2.16E-02*
	α1B β3 α2δ1	6.41E-01	2.34E-02*	3.92E-02*
	α1B β3 α2δ2	6.92E-01	3.28E-02*	1.02E-01
	α1B β3 α2δ3	3.58E-01	2.34E-02*	2.16E-02*
	α1B β4 α2δ1	6.90E-01	2.15E-02*	3.51E-02*
	α1B β4 α2δ2	7.66E-01	3.97E-02*	8.89E-02
	α1B β4 α2δ3	4.37E-01	2.15E-02*	1.92E-02*
Cav2.3	α1E β1 α2δ1	4.25E-01	5.15E-03*	1.47E-03*
	α1E β2 α2δ1	4.53E-02*	4.25E-04*	2.30E-07*
	α1E β3 α2δ1	4.25E-01	5.15E-03*	1.47E-03*
	α1E β4 α2δ1	4.82E-01	5.15E-03*	1.56E-03*
Cav3.1	α1G	2.23E-03*	1.28E-01	1.05E-03*
Cav3.2	α1H	4.82E-01	8.16E-01	3.08E-01
Cav3.3	α1I	7.31E-05*	3.85E-05*	1.64E-09*

^*:P-value < 0.05 after corrections.

Stage1: discovery phase; Stage 2: validation phase; BH: Benjamini Hochberg; SKAT: sequencing kernel association test.

SNP-set (Sequence) Kernel Association Test (SKAT)

SKAT was used to test for association between a set of genetic variants and dichotomous or quantitative phenotypes. It uses the logistic kernel-machine regression modeling framework. SKAT aggregates individual score test statistics of SNPs in a SNP set and computes SNP-set level P-values. SKAT can be used for common or/and rare variants (Ionita-Laza et al. 2013; Wu et al. 2010; Wu et al. 2011). In the current study, we focus on the common variants in line with the PGC schizophrenia study and used SKAT version 1.07 (Wu et al. 2010). The linear kernel with beta (p, 1.25), where p is the minor allele frequency of a SNP, was used. In our analysis, we carefully selected the cohort indicators and the first six principal components as covariates after comparing results including different number of principal components (three, six and ten) (Supplementary Table S1). At the same time, to overcome the issue of the large number of degrees of freedom, SKAT employs a test that adaptively estimates the degrees of freedom by accounting for correlation (LD) among the SNPs (Wu et al. 2010). In this study, a SNP set can be a collection of SNPs from a gene or several genes consisting of a heteromeric channel. The Benjamini Hochberg (BH) procedure was used to correct for multiple comparisons both in the Table 1 and Table 2 (Hochberg and Benjamini 1990; Wu et al. 2011).

Estimate schizophrenia heritability contributed by Ca_v Channels SNPs

Channels significantly associated with schizophrenia (Table 2) were selected. For each subtype of Ca_v channel, all of the auxiliary subunit (β, α₂δ, γ) genes contributing to a significant association with schizophrenia were grouped with each α₁ gene. The following gene lists Ca_v1.1 (CACNA1S, CACNA2D1, CACNB1, CACNG1), Ca_v1.2 (CACNA1C, CACNA2D1, CACNA2D2, CACNA2D3, CACNA2D4, CACNB1, CACNB2, CACNB3, CACNB4, CACNG1, CACNG2, CACNG3, CACNG4, CACNG5, CACNG6, CACNG7, CACNG8), Ca_v1.3 (CACNA1D, CACNA2D3, CACNB3, CACNB4), Ca_v2.1 (CACNA1A, CACNA2D1, CACNA2D3, CACNA2D4, CACNB1, CACNB4), Ca_v2.2 (CACNA1B, CACNA2D1, CACNA2D3, CACNB1, CACNB3, CACNB4), Ca_v2.3 (CACNA1E, CACNA2D1, CACNB1, CACNB2, CACNB3, CACNB4), Ca_v3.1 (CACNA1G), Ca_v3.3 (CACNA1I) were used to extract genotype-phenotype data for estimating chip heritability by using the linear mixed method BOLT-REML (Loh et al. 2015). The level of enrichment for association with schizophrenia was represented by the ratio of proportion of chip heritability (from each subtype of channel) in total heritability (33%) (Ripke et al. 2013) to the proportion of their SNPs in all SNPs (9423850 variants, minor allele frequency > 0.05) from the 1000 Genomes Project.

Results

Association of Ca_v genes with schizophrenia (gene level)

Two genes, CACNA1C and CACNA1I significantly associate with schizophrenia in the discovery cohort (corrected P < 0.05) and in the replication cohort (corrected P < 0.05) both according to the SNP-set (Sequence) Kernel Association Test (SKAT) method (Table 1) and univariate analysis (Supplementary Table S2). Within the combined sample (56,605 subjects) a further three genes were identified by the SKAT analysis: CACNA1E, CACNA1G and CACNB2. CACNA1C, CACNA1I and CACNB2 were previously reported, while CACNA1E and CACNA1G have not been reported as schizophrenia candidates (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014).

Association of Ca_v channels with schizophrenia (macromolecule level)

Macromolecule-level testing in the discovery cohort identified heteromers Ca_v1.2 (all possible subunits combinations), Ca_v2.3 (α_1E β₂ α₂δ₁), and monomers Ca_v3.1 (α_1G) and Ca_v3.3 (α_1I) as associated (corrected P < 0.05). All of them except Ca_v3.1 (α_1G) were replicated in the separate samples by SKAT analysis (Table 2). In the combined sample, heteromers Ca_v1.1 (α_1S β₁ α₂δ₁ γ₁), Ca_v1.2 (all possible subunits combinations), Ca_v1.3 (α_1D β₃ α₂δ₃, α_1D β₄ α₂δ₃), Ca_v2.1 (α_1A β₁ α₂δ₁, α_1A β₄ α₂δ₁, α_1A β₄ α₂δ₃, α_1A β₄ α₂δ₄), Ca_v2.2 (α_1B β₁ α₂δ₁, α_1B β₁ α₂δ₃, α_1B β₃ α₂δ₁, α_1B β₃ α₂δ₃, α_1B β₄ α₂δ₁, α_1B β₄ α₂δ₃), and Ca_v2.3 (α_1E β₁ α₂δ₁, α_1E β₂ α₂δ₁, α_1E β₃ α₂δ₁, α_1E β₄ α₂δ₁), and monomers Ca_v3.1 (α_1G) and Ca_v3.3 (α_1I) associate with the risk of schizophrenia (corrected P < 0.05) (Table 2).

Chip heritability of Ca_v channels

We estimate that 0.0567% (s.e. 0.0391%), 0.5051% (s.e. 0.1172%), 0.2453% (s.e. 0.0946%), 0.1788% (s.e. 0.0708%), 0.2578% (s.e. 0.0929%), 0.176% (s.e. 0.0658%), 0.0272% (s.e. 0.0316%) and 0.0569% (s.e. 0.0464%) of the variance in schizophrenia can be explained by Ca_v1.1, Ca_v1.2, Ca_v1.3, Ca_v2.1, Ca_v2.2, Ca_v2.3, Ca_v3.1 and Ca_v3.3 SNPs respectively (Fig. 2a). The Ca_v1.2 account for the largest amount of chip heritability (0.5051%, s.e. 0.1172%) and the Ca_v3.1 account for the least (0.0272%, s.e. 0.0316%). However, after accounting for the number of SNPs included in each Ca_v subtype, Ca_v3.1 and Ca_v3.3 show largest fold enrichment (39.83 and 36.51, respectively) (Fig. 2b). All tested subtypes of Ca_v channels show > 6-fold enrichment. The variance explained by each subtype of Ca_v channels is proportional to its number of SNPs (Supplementary Fig. S1). This is in line with the previous discovery that the larger the genomic region, the higher the proportion of chip heritability that can be accounted for (Yang et al. 2011).

Fig. 2.

Estimates of the schizophrenia variance explained by SNPs from each subtype of Ca_v channels. (a) Chip heritability of each significant subtype of Ca_v channel, (b) Fold enrichment of each significant subtype of Ca_v channel in schizophrenia. The fold enrichment is the ratio of the proportion of chip heritability (from each significant subtype of channel) in total heritability (33%) to the proportion of their SNPs in all SNPs (9,423,850 variants, minor allele frequency > 0.05) from 1000 Genomes Projects.

Robustness of the channel-based association

Ca_v channels that are significantly associated with schizophrenia reported by SKAT were also identified by another program MAGMA (de Leeuw et al. 2015) (Supplementary Table S5). However, MAGMA identified fewer channels at the discovery stage compared with SKAT (Table 2; Supplementary Table S5). But for the largest European dataset (49 sub-studies), MAGMA reports similar results with SKAT.

Discussion

In the current study we applied a macromolecule approach to a subsample of published schizophrenia GWAS (N = 56,605) and identified eight subtypes of Ca_v channels associated with schizophrenia, including the L-type Ca_v channels (Ca_v1.1, Ca_v1.2, Ca_v1.3), P-/Q-type Ca_v2.1, N-type Ca_v2.2, R-type Ca_v2.3, T-type channels (Ca_v3.1, Ca_v3.3). Only genes (CACNA1C, CACNB2 and CACNA1I) from Ca_v1.2 and Ca_v3.3 were implicated in the primary PGC analysis, which was based on a larger sample (N = 82,315) (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). In addition, we used another published statistical tool MAGMA to confirm our analysis. The results are highly consistent although the two programs are based on different assumptions and statistical models. It demonstrates that analyzing macromolecule subunit genes in aggregate is a complementary way to identify more genetic variants of schizophrenia compare to the traditional GWAS that treating each SNP separately.

The macromolecule subunits physically bind together to achieve their cellular functions, thus perturbations of any of their subunits may contribute to disease pathogenesis. In previous GWAS of schizophrenia, only a handful of channel subunits were implicated, perhaps due to the limited power of the massive univariate tests (Lichtenstein et al. 2009; Ripke et al. 2013; Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). To the best of our knowledge, only Askland and coworkers (Askland et al. 2012) have performed an association analysis of ion channels with schizophrenia, but the gene-sets defined in their study is a mixture of subunit-encoding genes from many ionic species and does not therefore correspond to a macromolecule existing in nature. In addition, it was tested in a much smaller sample. In order to test whether each functional Ca_v channel is associated with schizophrenia or not, we composed specific gene-set based on molecular structures of Ca_v channels (Buraei and Yang 2010; Catterall 1996; Davies et al. 2010; Schlick et al. 2010; Simms and Zamponi 2014). For each channel (macromolecule-based analysis), although the containing genes locate far away or even on different chromosomes, the encoding subunits are physically binding together in one functional unit to deal with flow of calcium ions. This macromolecule-based approach is different from grouping genes based on their functional catalogs or pathways since their products (proteins) interact directly or indirectly and they could not form a unique functional macromolecule. Our approach combining biological priors with GWAS data identified eight subtypes of Ca_v channels associated with the risk of schizophrenia. It is possible that the associations of whole channels with schizophrenia may be due to a highly associated component gene. This is likely the case for Ca_v1.2, where a few possible subunit combinations (for example, Ca_v1.2: α_1C β₁ α₂δ₂ that encoded by genes CACNA1C, CACNB1 and CACNA2D2) show their significance thanks to the α₁ subunit gene CACNA1C (Table 1; Supplementary Table S7) although most of the others are not. The significant associations of the other heteromultimeric channels may be not due to a single significant gene. For example, during the discovery and replication stages, the Ca_v2.3 channel (subunits encoded by CACNA1E, CACNB2 and CACNA2D1) was discovered and replicated by SKAT but none of their composing genes was identified at the gene-level test. The univariate analysis (minP SNP represents channel) could not identify this channel in small samples (discovery and replication stages), but the combined sample could confirm this finding when applying a macromolecule-based approach (Supplementary Table S3). None of the channels Ca_v1.1, Ca_v1.3, Ca_v2.1 and Ca_v2.2 subunit genes was identified in gene-level testing, but the channels show significant association with schizophrenia in the combined sample. These results indicate that subunit genes can collectively associate with disease susceptibility, even if individual genes do not exhibit significant association. It seems that analyzing channel SNPs as a set can capture the joint effect of multiple variants located on different chromosomes. Thus, genetic variants with weak or moderate effects could be identified when we combined them together based on biological knowledge of the macromolecule.

We also observed enrichment of heritability in significant Ca_v channels SNPs for schizophrenia and it may point to a major role of the inherited genetic variants in the risk of schizophrenia. These eight subtypes of Ca_v channels may provide more knowledge about the pathology of schizophrenia. Ca_v channels are the primary mediators of depolarization-induced calcium entry into neurons (Simms and Zamponi 2014). Calcium-dependent processes such as neurotransmitter release, neuronal gene transcription, and activation of calcium-dependent enzymes are of critical importance to brain function (Clapham 2007; Simms and Zamponi 2014). L-type Ca_v channels (Ca_v1.1, Ca_v1.2, Ca_v1.3) are involved in learning, memory and synaptic plasticity (Moosmang et al. 2005; White et al. 2008; Woodside et al. 2004). Mutations in CACNA1C, the gene encoding the α₁ subunit of Ca_v1.2, are responsible for Timothy syndrome, a multisystem disorder including cognitive impairment and autism spectrum disorder (Splawski et al. 2005; Splawski et al. 2004). SNPs located in CACNA1C are linked to development of schizophrenia, bipolar disorder and depression (Dao et al. 2010; Green et al. 2010; He et al. 2014). Data from mice and humans suggest an involvement of Ca_v1.3 channels in neurophysiological functions, in particular in the dopaminergic system (Simms and Zamponi 2014), which is involved in the pathology of schizophrenia (Brisch et al. 2014). Although, in humans, mutations in Ca_v1.1 have been linked to hypokalemic periodic paralysis (Ptáček et al. 1994) and malignant hyperthermia (Monnier et al. 1997), a pathway analysis for a set of calcium channel genes implicated CACNA1S (Ca_v1.1 channel α₁ subunit gene) as one of the 20 gene regions associated in the five psychiatric disorder meta-analysis (Cross-Disorder Group of the Psychiatric Genomics Consortium 2013). P-/Q-type channel Ca_v2.1 and N-type channel Ca_v2.2 play a role in neurotransmitter release at the presynaptic terminal and in neuronal integration in many neuronal types (Williams et al. 1992). R-type channel Ca_v2.3 are strongly expressed in cortex, hippocampus, striatum, amygdala and interpeduncular nucleus (Parajuli et al. 2012). The T-type channels (Ca_v3.1, Ca_v3.3) appear to play important roles in regulating neuronal excitability (Simms and Zamponi 2014). Although there is no direct evidence associating Ca_v2.1, Ca_v2.2, Ca_v2.3 and Ca_v3.1 with schizophrenia, due to their strong expression and wide distribution in the human brain, these four subtypes of Ca_v channels are likely involved in some aspects of schizophrenia pathology. A recent study of rare variants in schizophrenia demonstrated that a gene set containing 26 Ca_v genes yielded a large odds ratio of 8.4 (Purcell et al. 2014). Given the central role of Ca_v channels in regulating neurotransmitter release and neuronal gene transcription, the identified channels may represent convenient drug targets for novel therapeutics. Designing drugs for specific channels by targeting α₁ subunit, or designing more universal drugs for some channels by targeting shared ancillary subunits can improve efficiency of treatments. There are some Ca_v channels blockers in clinical use. A few L-type Ca_v channel antagonists such as verapamil and nifedipine which are used for hypertension, have been examined in clinical trials in schizophrenia (Lencz and Malhotra 2015). Revisiting the effect of existing agents on Ca_v channels or designing new drugs could be a high priority for new schizophrenia treatment development.

The genetic association test of macromolecules may also suggest candidates for non-additive interactions (epistasis) and improve polygenic predictions. In addition, while we only considered Ca_v channels, future work could consider other types of channels, such as potassium channels, sodium channels, and proton channels as interesting susceptibility candidates for schizophrenia and other psychiatric disorders.

The present findings illustrate the power of the macromolecule-based approach applied to schizophrenia, which identified eight subtypes of Ca_v channels associated with the disorder. The results highlight the combined role of different aspects of calcium signaling in schizophrenia pathophysiology, and suggest several new potential drug targets for development of novel therapeutics.

Schizophrenia Working Group of the Psychiatric Genomics Consortium

Stephan Ripke^1,2, Benjamin M. Neale^1,2,3,4, Aiden Corvin⁵, James T. R. Walters⁶, Kai-How Farh¹, Peter A. Holmans^6,7, Phil Lee^1,2,4, Brendan Bulik-Sullivan^1,2, David A. Collier^8,9, Hailiang Huang^1,3, Tune H. Pers^3,10,11, Ingrid Agartz^12,13,14, Esben Agerbo^15,16,17, Margot Albus¹⁸, Madeline Alexander¹⁹, Farooq Amin^20,21, Silviu A. Bacanu²², Martin Begemann²³, Richard A Belliveau Jr², Judit Bene^24,25, Sarah E. Bergen ^2,26, Elizabeth Bevilacqua², Tim B Bigdeli ²², Donald W. Black²⁷, Richard Bruggeman²⁸, Nancy G. Buccola²⁹, Randy L. Buckner^30,31,32, William Byerley³³, Wiepke Cahn³⁴, Guiqing Cai^35,36, Murray J. Cairns^39,120,170, Dominique Campion³⁷, Rita M. Cantor³⁸, Vaughan J. Carr^39,40, Noa Carrera⁶, Stanley V. Catts^39,41, Kimberly D. Chambert², Raymond C. K. Chan⁴², Ronald Y. L. Chen⁴³, Eric Y. H. Chen^43,44, Wei Cheng⁴⁵, Eric F. C. Cheung⁴⁶, Siow Ann Chong⁴⁷, C. Robert Cloninger⁴⁸, David Cohen⁴⁹, Nadine Cohen⁵⁰, Paul Cormican⁵, Nick Craddock^6,7, Benedicto Crespo-Facorro²¹⁰, James J. Crowley⁵¹, David Curtis^52,53, Michael Davidson⁵⁴, Kenneth L. Davis³⁶, Franziska Degenhardt^55,56, Jurgen Del Favero⁵⁷, Lynn E. DeLisi^128,129, Ditte Demontis^17,58,59, Dimitris Dikeos⁶⁰, Timothy Dinan⁶¹, Srdjan Djurovic^14,62, Gary Donohoe^5,63, Elodie Drapeau³⁶, Jubao Duan^64,65, Frank Dudbridge⁶⁶, Naser Durmishi⁶⁷, Peter Eichhammer⁶⁸, Johan Eriksson^69,70,71, Valentina Escott-Price⁶, Laurent Essioux⁷², Ayman H. Fanous^73,74,75,76, Martilias S. Farrell⁵¹, Josef Frank⁷⁷, Lude Franke⁷⁸, Robert Freedman⁷⁹, Nelson B. Freimer⁸⁰, Marion Friedl⁸¹, Joseph I. Friedman³⁶, Menachem Fromer^1,2,4,82, Giulio Genovese², Lyudmila Georgieva⁶, Elliot S. Gershon²⁰⁹, Ina Giegling^81,83, Paola Giusti-Rodríguez⁵¹, Stephanie Godard⁸⁴, Jacqueline I. Goldstein^1,3, Vera Golimbet⁸⁵, Srihari Gopal⁸⁶, Jacob Gratten⁸⁷, Lieuwe de Haan⁸⁸, Christian Hammer²³, Marian L. Hamshere⁶, Mark Hansen⁸⁹, Thomas Hansen^17,90, Vahram Haroutunian^36,91,92, Annette M. Hartmann⁸¹, Frans A. Henskens^39,93,94, Stefan Herms^55,56,95, Joel N. Hirschhorn^3,11,96, Per Hoffmann^55,56,95, Andrea Hofman^55,56, Mads V. Hollegaard⁹⁷, David M. Hougaard⁹⁷, Masashi Ikeda⁹⁸, Inge Joa⁹⁹, Antonio Julià¹⁰⁰, René S. Kahn³⁴, Luba Kalaydjieva^101,102, Sena Karachanak-Yankova¹⁰³, Juha Karjalainen⁷⁸, David Kavanagh⁶, Matthew C. Keller¹⁰⁴, Brian J. Kelly¹²⁰, James L. Kennedy^105,106,107, Andrey Khrunin¹⁰⁸, Yunjung Kim⁵¹, Janis Klovins¹⁰⁹, James A. Knowles¹¹⁰, Bettina Konte⁸¹, Vaidutis Kucinskas¹¹¹, Zita Ausrele Kucinskiene¹¹¹, Hana Kuzelova-Ptackova¹¹², Anna K. Kähler²⁶, Claudine Laurent^19,113, Jimmy Lee Chee Keong^47,114, S. Hong Lee⁸⁷, Sophie E. Legge⁶, Bernard Lerer¹¹⁵, Miaoxin Li^43,44,116 Tao Li¹¹⁷, Kung-Yee Liang¹¹⁸, Jeffrey Lieberman¹¹⁹, Svetlana Limborska¹⁰⁸, Carmel M. Loughland^39,120, Jan Lubinski¹²¹, Jouko Lönnqvist¹²², Milan Macek Jr¹¹², Patrik K. E. Magnusson²⁶, Brion S. Maher¹²³, Wolfgang Maier¹²⁴, Jacques Mallet¹²⁵, Sara Marsal¹⁰⁰, Manuel Mattheisen^17,58,59,126, Morten Mattingsdal^14,127, Robert W. McCarley^128,129, Colm McDonald¹³⁰, Andrew M. McIntosh^131,132, Sandra Meier⁷⁷, Carin J. Meijer⁸⁸, Bela Melegh^24,25, Ingrid Melle^14,133, Raquelle I. Mesholam-Gately^128,134, Andres Metspalu¹³⁵, Patricia T. Michie^39,136, Lili Milani¹³⁵, Vihra Milanova¹³⁷, Younes Mokrab⁸, Derek W. Morris^5,63, Ole Mors^17,58,138, Kieran C. Murphy¹³⁹, Robin M. Murray¹⁴⁰, Inez Myin-Germeys¹⁴¹, Bertram Müller-Myhsok^142,143,144, Mari Nelis¹³⁵, Igor Nenadic¹⁴⁵, Deborah A. Nertney¹⁴⁶, Gerald Nestadt¹⁴⁷, Kristin K. Nicodemus¹⁴⁸, Liene Nikitina-Zake¹⁰⁹, Laura Nisenbaum¹⁴⁹, Annelie Nordin¹⁵⁰, Eadbhard O’Callaghan¹⁵¹, Colm O’Dushlaine², F. Anthony O’Neill¹⁵², Sang-Yun Oh¹⁵³, Ann Olincy⁷⁹, Line Olsen^17,90, Jim Van Os^141,154, Psychosis Endophenotypes International Consortium¹⁵⁵, Christos Pantelis^39,156, George N. Papadimitriou⁶⁰, Sergi Papiol²³, Elena Parkhomenko³⁶, Michele T. Pato¹¹⁰, Tiina Paunio^157,158, Milica Pejovic-Milovancevic¹⁵⁹, Diana O. Perkins¹⁶⁰, Olli Pietiläinen^158,161, Jonathan Pimm⁵³, Andrew J. Pocklington⁶, John Powell¹⁴⁰, Alkes Price³,¹⁶², Ann E. Pulver¹⁴⁷, Shaun M. Purcell⁸², Digby Quested¹⁶³, Henrik B. Rasmussen^17,90, Abraham Reichenberg³⁶, Mark A. Reimers¹⁶⁴, Alexander L. Richards⁶, Joshua L. Roffman^30,32, Panos Roussos^82,165, Douglas M. Ruderfer^6,82, Veikko Salomaa⁷¹, Alan R. Sanders^64,65, Ulrich Schall^39,120, Christian R. Schubert¹⁶⁶, Thomas G. Schulze^77,167, Sibylle G. Schwab¹⁶⁸, Edward M. Scolnick², Rodney J. Scott^39,169,170, Larry J. Seidman^128,134, Jianxin Shi¹⁷¹, Engilbert Sigurdsson¹⁷², Teimuraz Silagadze¹⁷³, Jeremy M. Silverman^36,174, Kang Sim⁴⁷, Petr Slominsky¹⁰⁸, Jordan W. Smoller^2,4, Hon-Cheong So⁴³, Chris C. A. Spencer¹⁷⁵, Eli A. Stahl^3,82, Hreinn Stefansson¹⁷⁶, Stacy Steinberg¹⁷⁶, Elisabeth Stogmann¹⁷⁷, Richard E. Straub¹⁷⁸, Eric Strengman^179,34, Jana Strohmaier⁷⁷, T. Scott Stroup¹¹⁹, Mythily Subramaniam⁴⁷, Jaana Suvisaari¹²², Dragan M. Svrakic⁴⁸, Jin P. Szatkiewicz⁵¹, Erik Söderman¹², Srinivas Thirumalai¹⁸⁰, Draga Toncheva¹⁰³, Paul A. Tooney^39,120,170, Sarah Tosato¹⁸¹, Juha Veijola^182,183, John Waddington¹⁸⁴, Dermot Walsh¹⁸⁵, Dai Wang⁸⁶, Qiang Wang¹¹⁷, Bradley T. Webb²², Mark Weiser⁵⁴, Dieter B. Wildenauer¹⁸⁶, Nigel M. Williams⁶, Stephanie Williams⁵¹, Stephanie H. Witt⁷⁷, Aaron R. Wolen¹⁶⁴, Emily H. M. Wong⁴³, Brandon K. Wormley²², Jing Qin Wu^39,170, Hualin Simon Xi¹⁸⁷, Clement C. Zai^105,106, Xuebin Zheng¹⁸⁸, Fritz Zimprich¹⁷⁷, Naomi R. Wray⁸⁷, Kari Stefansson¹⁷⁶, Peter M. Visscher⁸⁷, Wellcome Trust Case-Control Consortium 2¹⁸⁹, Rolf Adolfsson¹⁵⁰, Ole A. Andreassen^14,133, Douglas H. R. Blackwood¹³², Elvira Bramon¹⁹⁰, Joseph D. Buxbaum^35,36,91,191, Anders D. Børglum^17,58,59,138, Sven Cichon^55,56,95,192, Ariel Darvasi¹⁹³, Enrico Domenici¹⁹⁴, Hannelore Ehrenreich²³, Tõnu Esko^3,11,96,135, Pablo V. Gejman^64,65, Michael Gill⁵, Hugh Gurling⁵³, Christina M. Hultman²⁶, Nakao Iwata⁹⁸, Assen V. Jablensky^{39,102,186,195}, Erik G. Jönsson^12,14, Kenneth S. Kendler¹⁹⁶, George Kirov⁶, Jo Knight^105,106,107, Todd Lencz^197,198,199, Douglas F. Levinson¹⁹, Qingqin S. Li⁸⁶, Jianjun Liu^188,200, Anil K. Malhotra^197,198,199, Steven A. McCarroll^2,96, Andrew McQuillin⁵³, Jennifer L. Moran², Preben B. Mortensen^15,16,17, Bryan J. Mowry^87,201, Markus M. Nöthen^55,56, Roel A. Ophoff^38,80,34, Michael J. Owen^6,7, Aarno Palotie^2,4,161, Carlos N. Pato¹¹⁰, Tracey L. Petryshen^2,128,202, Danielle Posthuma^203,204,205, Marcella Rietschel⁷⁷, Brien P. Riley¹⁹⁶, Dan Rujescu^81,83, Pak C. Sham^43,44,116 Pamela Sklar^82,91,165, David St Clair²⁰⁶, Daniel R. Weinberger^178,207, Jens R. Wendland¹⁶⁶, Thomas Werge^17,90,208, Mark J. Daly^1,2,3, Patrick F. Sullivan^26,51,160 & Michael C. O’Donovan^6,7

¹Analytic and Translational Genetics Unit, Massachusetts General Hospital, Boston, Massachusetts 02114, USA.

²Stanley Center for Psychiatric Research, Broad Institute of MIT and Harvard, Cambridge, Massachusetts 02142, USA.

³Medical and Population Genetics Program, Broad Institute of MIT and Harvard, Cambridge, Massachusetts 02142, USA.

⁴Psychiatric and Neurodevelopmental Genetics Unit, Massachusetts General Hospital, Boston, Massachusetts 02114, USA.

⁵Neuropsychiatric Genetics Research Group, Department of Psychiatry, Trinity College Dublin, Dublin 8, Ireland.

⁶MRC Centre for Neuropsychiatric Genetics and Genomics, Institute of Psychological Medicine and Clinical Neurosciences, School of Medicine, Cardiff University, Cardiff, CF24 4HQ, UK.

⁷National Centre for Mental Health, Cardiff University, Cardiff, CF24 4HQ, UK.

⁸Eli Lilly and Company Limited, Erl Wood Manor, Sunninghill Road, Windlesham, Surrey, GU20 6PH, UK.

⁹Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, King’s College London, London, SE5 8AF, UK.

¹⁰Center for Biological Sequence Analysis, Department of Systems Biology, Technical University of Denmark, DK-2800, Denmark.

¹¹Division of Endocrinology and Center for Basic and Translational Obesity Research, Boston Children’s Hospital, Boston, Massachusetts, 02115USA.

¹²Department of Clinical Neuroscience, Psychiatry Section, Karolinska Institutet, SE-17176 Stockholm, Sweden.

¹³Department of Psychiatry, Diakonhjemmet Hospital, 0319 Oslo, Norway.

¹⁴NORMENT, KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, University of Oslo, 0424 Oslo, Norway.

¹⁵Centre for Integrative Register-based Research, CIRRAU, Aarhus University, DK-8210 Aarhus, Denmark.

¹⁶National Centre for Register-based Research, Aarhus University, DK-8210 Aarhus, Denmark.

¹⁷The Lundbeck Foundation Initiative for Integrative Psychiatric Research, iPSYCH, Denmark.

¹⁸State Mental Hospital, 85540 Haar, Germany.

¹⁹Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, California 94305, USA.

²⁰Department of Psychiatry and Behavioral Sciences, Atlanta Veterans Affairs Medical Center, Atlanta, Georgia 30033, USA.

²¹Department of Psychiatry and Behavioral Sciences, Emory University, Atlanta Georgia 30322, USA.

²²Virginia Institute for Psychiatric and Behavioral Genetics, Department of Psychiatry, Virginia Commonwealth University, Richmond, Virginia 23298, USA.

²³Clinical Neuroscience, Max Planck Institute of Experimental Medicine, Göttingen 37075, Germany.

²⁴Department of Medical Genetics, University of Pécs, Pécs H-7624, Hungary.

²⁵Szentagothai Research Center, University of Pécs, Pécs H-7624, Hungary.

²⁶Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm SE-17177, Sweden.

²⁷Department of Psychiatry, University of Iowa Carver College of Medicine, Iowa City, Iowa 52242, USA.

²⁸University Medical Center Groningen, Department of Psychiatry, University of Groningen NL-9700 RB, The Netherlands.

²⁹School of Nursing, Louisiana State University Health Sciences Center, New Orleans, Louisiana 70112, USA.

³⁰Athinoula A. Martinos Center, Massachusetts General Hospital, Boston, Massachusetts 02129, USA.

³¹Center for Brain Science, Harvard University, Cambridge, Massachusetts, 02138 USA.

³²Department of Psychiatry, Massachusetts General Hospital, Boston, Massachusetts, 02114 USA.

³³Department of Psychiatry, University of California at San Francisco, San Francisco, California, 94143 USA.

³⁴University Medical Center Utrecht, Department of Psychiatry, Rudolf Magnus Institute of Neuroscience, 3584 Utrecht, The Netherlands.

³⁵Department of Human Genetics, Icahn School of Medicine at Mount Sinai, New York, New York 10029 USA.

³⁶Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, New York 10029 USA.

³⁷Centre Hospitalier du Rouvray and INSERM U1079 Faculty of Medicine, 76301 Rouen, France.

³⁸Department of Human Genetics, David Geffen School of Medicine, University of California, Los Angeles, California 90095, USA.

³⁹Schizophrenia Research Institute, Sydney NSW 2010, Australia.

⁴⁰School of Psychiatry, University of New South Wales, Sydney NSW 2031, Australia.

⁴¹Royal Brisbane and Women’s Hospital, University of Queensland, Brisbane, St Lucia QLD 4072, Australia.

⁴²Institute of Psychology, Chinese Academy of Science, Beijing 100101, China.

⁴³Department of Psychiatry, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China.

⁴⁴State Key Laboratory for Brain and Cognitive Sciences, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China.

⁴⁵Department of Computer Science, University of North Carolina, Chapel Hill, North Carolina 27514, USA.

⁴⁶Castle Peak Hospital, Hong Kong, China.

⁴⁷Institute of Mental Health, Singapore 539747, Singapore.

⁴⁸Department of Psychiatry, Washington University, St. Louis, Missouri 63110, USA.

⁴⁹Department of Child and Adolescent Psychiatry, Assistance Publique Hopitaux de Paris, Pierre and Marie Curie Faculty of Medicine and Institute for Intelligent Systems and Robotics, Paris, 75013, France.

⁵⁰Blue Note Biosciences, Princeton, New Jersey 08540, USA

⁵¹Department of Genetics, University of North Carolina, Chapel Hill, North Carolina 27599–7264, USA.

⁵²Department of Psychological Medicine, Queen Mary University of London, London E1 1BB, UK.

⁵³Molecular Psychiatry Laboratory, Division of Psychiatry, University College London, London WC1E 6JJ, UK.

⁵⁴Sheba Medical Center, Tel Hashomer 52621, Israel.

⁵⁵Department of Genomics, Life and Brain Center, D-53127 Bonn, Germany.

⁵⁶Institute of Human Genetics, University of Bonn, D-53127 Bonn, Germany.

⁵⁷Applied Molecular Genomics Unit, VIB Department of Molecular Genetics, University of Antwerp, B-2610 Antwerp, Belgium.

⁵⁸Centre for Integrative Sequencing, iSEQ, Aarhus University, DK-8000 Aarhus C, Denmark.

⁵⁹Department of Biomedicine, Aarhus University, DK-8000 Aarhus C, Denmark.

⁶⁰First Department of Psychiatry, University of Athens Medical School, Athens 11528, Greece.

⁶¹Department of Psychiatry, University College Cork, Co. Cork, Ireland.

⁶²Department of Medical Genetics, Oslo University Hospital, 0424 Oslo, Norway.

⁶³Cognitive Genetics and Therapy Group, School of Psychology and Discipline of Biochemistry, National University of Ireland Galway, Co. Galway, Ireland.

⁶⁴Department of Psychiatry and Behavioral Neuroscience, University of Chicago, Chicago, Illinois 60637, USA.

⁶⁵Department of Psychiatry and Behavioral Sciences, NorthShore University HealthSystem, Evanston, Illinois 60201, USA.

⁶⁶Department of Non-Communicable Disease Epidemiology, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK.

⁶⁷Department of Child and Adolescent Psychiatry, University Clinic of Psychiatry, Skopje 1000, Republic of Macedonia.

⁶⁸Department of Psychiatry, University of Regensburg, 93053 Regensburg, Germany.

⁶⁹Department of General Practice, Helsinki University Central Hospital, University of Helsinki P.O. Box 20, Tukholmankatu 8 B, FI-00014, Helsinki, Finland

⁷⁰Folkhälsan Research Center, Helsinki, Finland, Biomedicum Helsinki 1, Haartmaninkatu 8, FI-00290, Helsinki, Finland.

⁷¹National Institute for Health and Welfare, P.O. BOX 30, FI-00271 Helsinki, Finland.

⁷²Translational Technologies and Bioinformatics, Pharma Research and Early Development, F. Hoffman-La Roche, CH-4070 Basel, Switzerland.

⁷³Department of Psychiatry, Georgetown University School of Medicine, Washington DC 20057, USA.

⁷⁴Department of Psychiatry, Keck School of Medicine of the University of Southern California, Los Angeles, California 90033, USA.

⁷⁵Department of Psychiatry, Virginia Commonwealth University School of Medicine, Richmond, Virginia 23298, USA.

⁷⁶Mental Health Service Line, Washington VA Medical Center, Washington DC 20422, USA.

⁷⁷Department of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, Medical Faculty Mannheim, University of Heidelberg, Heidelberg, D-68159 Mannheim, Germany.

⁷⁸Department of Genetics, University of Groningen, University Medical Centre Groningen, 9700 RB Groningen, The Netherlands.

⁷⁹Department of Psychiatry, University of Colorado Denver, Aurora, Colorado 80045, USA.

⁸⁰Center for Neurobehavioral Genetics, Semel Institute for Neuroscience and Human Behavior, University of California, Los Angeles, California 90095, USA.

⁸¹Department of Psychiatry, University of Halle, 06112 Halle, Germany.

⁸²Division of Psychiatric Genomics, Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA.

⁸³Department of Psychiatry, University of Munich, 80336, Munich, Germany.

⁸⁴Departments of Psychiatry and Human and Molecular Genetics, INSERM, Institut de Myologie, Hôpital de la Pitiè-Salpêtrière, Paris, 75013, France.

⁸⁵Mental Health Research Centre, Russian Academy of Medical Sciences, 115522 Moscow, Russia.

⁸⁶Neuroscience Therapeutic Area, Janssen Research and Development, Raritan, New Jersey 08869, USA.

⁸⁷Queensland Brain Institute, The University of Queensland, Brisbane, Queensland, QLD 4072, Australia.

⁸⁸Academic Medical Centre University of Amsterdam, Department of Psychiatry, 1105 AZ Amsterdam, The Netherlands.

⁸⁹Illumina, La Jolla, California, California 92122, USA.

⁹⁰Institute of Biological Psychiatry, Mental Health Centre Sct. Hans, Mental Health Services Copenhagen, DK-4000, Denmark.

⁹¹Friedman Brain Institute, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA.

⁹²J. J. Peters VA Medical Center, Bronx, New York, New York 10468, USA.

⁹³Priority Research Centre for Health Behaviour, University of Newcastle, Newcastle NSW 2308, Australia.

⁹⁴School of Electrical Engineering and Computer Science, University of Newcastle, Newcastle NSW 2308, Australia.

⁹⁵Division of Medical Genetics, Department of Biomedicine, University of Basel, Basel, CH-4058, Switzerland.

⁹⁶Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115, USA.

⁹⁷Section of Neonatal Screening and Hormones, Department of Clinical Biochemistry, Immunology and Genetics, Statens Serum Institut, Copenhagen, DK-2300, Denmark.

⁹⁸Department of Psychiatry, Fujita Health University School of Medicine, Toyoake, Aichi, 470–1192, Japan.

⁹⁹Regional Centre for Clinical Research in Psychosis, Department of Psychiatry, Stavanger University Hospital, 4011 Stavanger, Norway.

¹⁰⁰Rheumatology Research Group, Vall d’Hebron Research Institute, Barcelona, 08035, Spain.

¹⁰¹Centre for Medical Research, The University of Western Australia, Perth, WA 6009, Australia.

¹⁰²The Perkins Institute for Medical Research, The University of Western Australia, Perth, WA 6009, Australia.

¹⁰³Department of Medical Genetics, Medical University, Sofia1431, Bulgaria.

¹⁰⁴Department of Psychology, University of Colorado Boulder, Boulder, Colorado 80309, USA.

¹⁰⁵Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, Ontario, M5T 1R8, Canada.

¹⁰⁶Department of Psychiatry, University of Toronto, Toronto, Ontario, M5T 1R8, Canada.

¹⁰⁷Institute of Medical Science, University of Toronto, Toronto, Ontario, M5S 1A8, Canada.

¹⁰⁸Institute of Molecular Genetics, Russian Academy of Sciences, Moscow123182, Russia.

¹⁰⁹Latvian Biomedical Research and Study Centre, Riga, LV-1067, Latvia.

¹¹⁰Department of Psychiatry and Zilkha Neurogenetics Institute, Keck School of Medicine at University of Southern California, Los Angeles, California 90089, USA.

¹¹¹Faculty of Medicine, Vilnius University, LT-01513 Vilnius, Lithuania.

¹¹² Department of Biology and Medical Genetics, 2nd Faculty of Medicine and University Hospital Motol, 150 06 Prague, Czech Republic.

¹¹³ Department of Child and Adolescent Psychiatry, Pierre and Marie Curie Faculty of Medicine, Paris 75013, France.

¹¹⁴Duke-NUS Graduate Medical School, Singapore 169857, Singapore.

¹¹⁵Department of Psychiatry, Hadassah-Hebrew University Medical Center, Jerusalem 91120, Israel.

¹¹⁶Centre for Genomic Sciences, The University of Hong Kong, Hong Kong, China.

¹¹⁷Mental Health Centre and Psychiatric Laboratory, West China Hospital, Sichuan University, Chengdu, 610041, Sichuan, China.

¹¹⁸Department of Biostatistics, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland 21205, USA.

¹¹⁹Department of Psychiatry, Columbia University, New York, New York 10032, USA.

¹²⁰Priority Centre for Translational Neuroscience and Mental Health, University of Newcastle, Newcastle NSW 2300, Australia.

¹²¹Department of Genetics and Pathology, International Hereditary Cancer Center, Pomeranian Medical University in Szczecin, 70–453 Szczecin, Poland.

¹²²Department of Mental Health and Substance Abuse Services; National Institute for Health and Welfare, P.O. BOX 30, FI-00271 Helsinki, Finland

¹²³Department of Mental Health, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, Maryland 21205, USA.

¹²⁴Department of Psychiatry, University of Bonn, D-53127 Bonn, Germany.

¹²⁵Centre National de la Recherche Scientifique, Laboratoire de Génétique Moléculaire de la Neurotransmission et des Processus Neurodégénératifs, Hôpital de la Pitié Salpêtrière, 75013, Paris, France.

¹²⁶Department of Genomics Mathematics, University of Bonn, D-53127 Bonn, Germany.

¹²⁷Research Unit, Sørlandet Hospital, 4604 Kristiansand, Norway.

¹²⁸Department of Psychiatry, Harvard Medical School, Boston, Massachusetts 02115, USA.

¹²⁹VA Boston Health Care System, Brockton, Massachusetts 02301, USA.

¹³⁰Department of Psychiatry, National University of Ireland Galway, Co. Galway, Ireland.

¹³¹Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, Edinburgh EH16 4SB, UK.

¹³²Division of Psychiatry, University of Edinburgh, Edinburgh EH16 4SB, UK.

¹³³Division of Mental Health and Addiction, Oslo University Hospital, 0424 Oslo, Norway.

¹³⁴Massachusetts Mental Health Center Public Psychiatry Division of the Beth Israel Deaconess Medical Center, Boston, Massachusetts 02114, USA.

¹³⁵Estonian Genome Center, University of Tartu, Tartu 50090, Estonia.

¹³⁶School of Psychology, University of Newcastle, Newcastle NSW 2308, Australia.

¹³⁷First Psychiatric Clinic, Medical University, Sofia 1431, Bulgaria.

¹³⁸Department P, Aarhus University Hospital, DK-8240 Risskov, Denmark.

¹³⁹Department of Psychiatry, Royal College of Surgeons in Ireland, Dublin 2, Ireland.

¹⁴⁰King’s College London, London SE5 8AF, UK.

¹⁴¹Maastricht University Medical Centre, South Limburg Mental Health Research and Teaching Network, EURON, 6229 HX Maastricht, The Netherlands.

¹⁴²Institute of Translational Medicine, University of Liverpool, Liverpool L69 3BX, UK.

¹⁴³Max Planck Institute of Psychiatry, 80336 Munich, Germany.

¹⁴⁴Munich Cluster for Systems Neurology (SyNergy), 80336 Munich, Germany.

¹⁴⁵Department of Psychiatry and Psychotherapy, Jena University Hospital, 07743 Jena, Germany.

¹⁴⁶Department of Psychiatry, Queensland Brain Institute and Queensland Centre for Mental Health Research, University of Queensland, Brisbane, Queensland, St Lucia QLD 4072, Australia.

¹⁴⁷Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205, USA.

¹⁴⁸Department of Psychiatry, Trinity College Dublin, Dublin 2, Ireland.

¹⁴⁹Eli Lilly and Company, Lilly Corporate Center, Indianapolis, 46285 Indiana, USA.

¹⁵⁰Department of Clinical Sciences, Psychiatry, Umeå University, SE-901 87 Umeå, Sweden.

¹⁵¹DETECT Early Intervention Service for Psychosis, Blackrock, Co. Dublin, Ireland.

¹⁵²Centre for Public Health, Institute of Clinical Sciences, Queen’s University Belfast, Belfast BT12 6AB, UK.

¹⁵³Lawrence Berkeley National Laboratory, University of California at Berkeley, Berkeley, California 94720, USA.

¹⁵⁴Institute of Psychiatry, King’s College London, London SE5 8AF, UK.

¹⁵⁵A list of authors and affiliations appear in the Supplementary Information.

¹⁵⁶Melbourne Neuropsychiatry Centre, University of Melbourne & Melbourne Health, Melbourne, Vic 3053, Australia.

¹⁵⁷Department of Psychiatry, University of Helsinki, P.O. Box 590, FI-00029 HUS, Helsinki, Finland.

¹⁵⁸Public Health Genomics Unit, National Institute for Health and Welfare, P.O. BOX 30, FI-00271 Helsinki, Finland

¹⁵⁹Medical Faculty, University of Belgrade, 11000 Belgrade, Serbia.

¹⁶⁰Department of Psychiatry, University of North Carolina, Chapel Hill, North Carolina 27599–7160, USA.

¹⁶¹Institute for Molecular Medicine Finland, FIMM, University of Helsinki, P.O. Box 20 FI-00014, Helsinki, Finland

¹⁶²Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts 02115, USA.

¹⁶³Department of Psychiatry, University of Oxford, Oxford, OX3 7JX, UK.

¹⁶⁴Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, Virginia 23298, USA.

¹⁶⁵Institute for Multiscale Biology, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA.

¹⁶⁶PharmaTherapeutics Clinical Research, Pfizer Worldwide Research and Development, Cambridge, Massachusetts 02139, USA.

¹⁶⁷Department of Psychiatry and Psychotherapy, University of Gottingen, 37073 Göttingen, Germany.

¹⁶⁸Psychiatry and Psychotherapy Clinic, University of Erlangen, 91054 Erlangen, Germany.

¹⁶⁹Hunter New England Health Service, Newcastle NSW 2308, Australia.

¹⁷⁰School of Biomedical Sciences and Pharmacy, University of Newcastle, Callaghan NSW 2308, Australia.

¹⁷¹Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, Maryland 20892, USA.

¹⁷²University of Iceland, Landspitali, National University Hospital, 101 Reykjavik, Iceland.

¹⁷³Department of Psychiatry and Drug Addiction, Tbilisi State Medical University (TSMU), N33, 0177 Tbilisi, Georgia.

¹⁷⁴Research and Development, Bronx Veterans Affairs Medical Center, New York, New York 10468, USA.

¹⁷⁵Wellcome Trust Centre for Human Genetics, Oxford, OX3 7BN, UK.

¹⁷⁶deCODE Genetics, 101 Reykjavik, Iceland.

¹⁷⁷Department of Clinical Neurology, Medical University of Vienna, 1090 Wien, Austria.

¹⁷⁸Lieber Institute for Brain Development, Baltimore, Maryland 21205, USA.

¹⁷⁹Department of Medical Genetics, University Medical Centre Utrecht, Universiteitsweg 100, 3584 CG, Utrecht, The Netherlands.

¹⁸⁰Berkshire Healthcare NHS Foundation Trust, Bracknell RG12 1BQ, UK.

¹⁸¹Section of Psychiatry, University of Verona, 37134 Verona, Italy.

¹⁸²Department of Psychiatry, University of Oulu, P.O. BOX 5000, 90014, Finland

¹⁸³University Hospital of Oulu, P.O.BOX 20, 90029 OYS, Finland.

¹⁸⁴Molecular and Cellular Therapeutics, Royal College of Surgeons in Ireland, Dublin 2, Ireland.

¹⁸⁵Health Research Board, Dublin 2, Ireland.

¹⁸⁶School of Psychiatry and Clinical Neurosciences, The University of Western Australia, Perth WA6009, Australia.

¹⁸⁷Computational Sciences CoE, Pfizer Worldwide Research and Development, Cambridge, Massachusetts 02139, USA.

¹⁸⁸Human Genetics, Genome Institute of Singapore, A*STAR, Singapore 138672, Singapore.

¹⁸⁹A list of authors and affiliations appear in the Supplementary Information.

¹⁹⁰University College London, London WC1E 6BT, UK.

¹⁹¹Department of Neuroscience, Icahn School of Medicine at Mount Sinai, New York, New York 10029, USA.

¹⁹²Institute of Neuroscience and Medicine (INM-1), Research Center Juelich, 52428 Juelich, Germany.

¹⁹³Department of Genetics, The Hebrew University of Jerusalem, 91905 Jerusalem, Israel.

¹⁹⁴Neuroscience Discovery and Translational Area, Pharma Research and Early Development, F. Hoffman-La Roche, CH-4070 Basel, Switzerland.

¹⁹⁵Centre for Clinical Research in Neuropsychiatry, School of Psychiatry and Clinical Neurosciences, The University of Western Australia, Medical Research Foundation Building, Perth WA 6000, Australia.

¹⁹⁶Virginia Institute for Psychiatric and Behavioral Genetics, Departments of Psychiatry and Human and Molecular Genetics, Virginia Commonwealth University, Richmond, Virginia 23298, USA.

¹⁹⁷The Feinstein Institute for Medical Research, Manhasset, New York, 11030 USA.

¹⁹⁸The Hofstra NS-LIJ School of Medicine, Hempstead, New York, 11549 USA.

¹⁹⁹The Zucker Hillside Hospital, Glen Oaks, New York,11004 USA.

²⁰⁰Saw Swee Hock School of Public Health, National University of Singapore, Singapore 117597, Singapore.

²⁰¹Queensland Centre for Mental Health Research, University of Queensland, Brisbane 4076, Queensland, Australia.

²⁰²Center for Human Genetic Research and Department of Psychiatry, Massachusetts General Hospital, Boston, Massachusetts 02114, USA.

²⁰³Department of Child and Adolescent Psychiatry, Erasmus University Medical Centre, Rotterdam 3000, The Netherlands.

²⁰⁴Department of Complex Trait Genetics, Neuroscience Campus Amsterdam, VU University Medical Center Amsterdam, Amsterdam 1081, The Netherlands.

²⁰⁵Department of Functional Genomics, Center for Neurogenomics and Cognitive Research, Neuroscience Campus Amsterdam, VU University, Amsterdam 1081, The Netherlands.

²⁰⁶University of Aberdeen, Institute of Medical Sciences, Aberdeen, AB25 2ZD, UK.

²⁰⁷Departments of Psychiatry, Neurology, Neuroscience and Institute of Genetic Medicine, Johns Hopkins School of Medicine, Baltimore, Maryland 21205, USA.

²⁰⁸Department of Clinical Medicine, University of Copenhagen, Copenhagen 2200, Denmark.

²⁰⁹Departments of Psychiatry and Human Genetics, University of Chicago, Chicago, Illinois 60637, USA.

²¹⁰University Hospital Marqués de Valdecilla, Instituto de Formación e Investigación Marqués de Valdecilla, University of Cantabria, E‐39008 Santander, Spain