Genome-wide association study meta-analysis of European and Asian-ancestry samples identifies three novel loci associated with bipolar disorder

  • A Corrigendum to this article was published on 24 January 2013


Meta-analyses of bipolar disorder (BD) genome-wide association studies (GWAS) have identified several genome-wide significant signals in European-ancestry samples, but so far account for little of the inherited risk. We performed a meta-analysis of 750 000 high-quality genetic markers on a combined sample of 14 000 subjects of European and Asian-ancestry (phase I). The most significant findings were further tested in an extended sample of 17 700 cases and controls (phase II). The results suggest novel association findings near the genes TRANK1 (LBA1), LMAN2L and PTGFR. In phase I, the most significant single nucleotide polymorphism (SNP), rs9834970 near TRANK1, was significant at the P=2.4 × 10−11 level, with no heterogeneity. Supportive evidence for prior association findings near ANK3 and a locus on chromosome 3p21.1 was also observed. The phase II results were similar, although the heterogeneity test became significant for several SNPs. On the basis of these results and other established risk loci, we used the method developed by Park et al. to estimate the number, and the effect size distribution, of BD risk loci that could still be found by GWAS methods. We estimate that >63 000 case–control samples would be needed to identify the 105 BD risk loci discoverable by GWAS, and that these will together explain <6% of the inherited risk. These results support previous GWAS findings and identify three new candidate genes for BD. Further studies are needed to replicate these findings and may potentially lead to identification of functional variants. Sample size will remain a limiting factor in the discovery of common alleles associated with BD.


The genetic basis of bipolar disorder (BD) is still largely unknown despite robust evidence of high heritability estimated by previous twin studies.1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 With a lifetime prevalence worldwide between 0.5 and 1.5%, BD is characterized clinically by often disabling fluctuations of mood and behavior, commonly developing in late adolescence to early adulthood. Although the pathogenesis of BD remains unclear, genome-wide association studies (GWAS) have so far identified and replicated a few risk loci (near the genes DGKH, ANK3 and CACNA1C),12, 13, 14, 15, 16 along with a locus on chromosome 3p21.1 that harbors a number of genes.17 Together, these loci account for little of the BD heritability, suggesting that additional risk loci remain undiscovered. The total BD GWAS sample size studied, so far, remains low compared with many other common traits studied, such as type 2 diabetes, height, serum lipids, colorectal cancer and rheumatoid arthritis.18, 19, 20, 21, 22, 23 Some of the missing heritability may be explained by additional risk loci that can only be identified in larger sample sizes.24

Psychiatric disorders such as BD pose statistical challenges when it comes to very large sample sizes. Phenotyping by direct diagnostic evaluation is expensive. Reliance on physician or hospital-assigned diagnoses can save money, but introduces potential biases, like changing diagnostic-criteria, which can be difficult to correct.25 Increasing sample size by combining data across studies, can be fruitful. Meta-analysis is an efficient and largely unbiased way to increase effective sample size by systematically combining association signals across studies. As most common genetic variation is ancient and widespread, some risk alleles may be shared across continental populations. It is possible to use meta-analysis to combine study samples of differing ancestry, as long as appropriate ancestry-matched controls are used within each study.26, 27, 28

In this study, we have sought to identify novel risk alleles for BD by meta-analysis of world-wide BD GWAS, comprising case–control samples of both European and Asian ancestry. The combined sample size of 17 656 is the largest so far in BD, to our knowledge. The results suggest significant novel association signals near the genes TRANK1 (LBA1), LMAN2L and PTGFR, and provide supportive evidence for the previously reported association signals near ANK3 and within the 3p21.1 locus. Largely consistent signals were observed in both the European ancestry and Asian-ancestry samples. Based on these findings and discoveries to date, we also present a GWAS discovery trajectory for BD.

Materials and methods

Study samples

The samples used in the meta-analysis have been described previously, and details are provided in Table 1 and Supplementary Table 1.12, 15, 17, 29, 30, 31, 32 For phase I, we obtained five European and one Asian-ancestry sample, totaling 14 000 cases and controls. The most significant hits (P<4 × 10−3) were tested in an extended sample that included phase I plus two independent European ancestry samples (3800 cases/controls). We refer to this as the phase II sample.

Table 1 Descriptive statistics for the samples analyzed


Genotype data from the NIMH-GAIN, German and TGEN samples were used to impute data on 2.1 million HapMap phase2 single-nucleotide polymorphism (SNPs), by use of the program Markov Chain Haplotyping (MACH version1.0; MACH uses Markov chain haplotyping to resolve haplotypes, and thereby missing genotypes, from observed genotypes in unrelated individuals. We used the ‘greedy’ algorithm, as recommended by the authors. SNPs flagged as having different alleles than in HapMap CEU or as monomorphic were reviewed, after which they were recoded for the reverse-strand (flipped) or dropped. SNPs flagged for allele frequencies markedly different from HapMap CEU were also reviewed. Palindromic SNPs whose allele frequencies were consistent with reversed coding were flipped. SNPs with unexpected allele frequencies were dropped. PLINK (version1.4) was used to flip and drop SNPs.34 After all allele-coding, monomorphism and palindrome issues were resolved, imputation was run again. SNPs in the result files were dropped if the minor allele frequency (MAF) in cases or controls was <0.05, or if the error rate (in the .erate output file) was >0.01. The imputed data were then formatted into PLINK binaries for analysis. Supplementary Table 1 provides detailed description regarding genotyping and imputation for the Taiwan, Wellcome Trust Case–Control Consortium, STEP-BD,35 FondaMental Bipolar and GlaxoSmithKline (GSK) samples.


PLINK output (assoc) files were modified with columns for direction-of-association, sample size and strand. For most samples, sample size equaled the sum of cases and controls included in the final analysis, after the quality-control steps were complete. For STEP-BD, sample size was set to equal the number of cases plus controls that did not overlap with those in the NIMH-GAIN or TGEN. This was done to avoid over-weighting the results from the NIMH-control sample, overlapping portions of which were included in both the NIMH-GAIN, TGEN and STEP-BD. Modified files were loaded into METAL (November 2010 version), then processed using the GENOMICCONTROL option, which applies a genomic control36 correction in samples where the genomic inflation factor is >1.0. METAL weights each sample based on the square root of the sample size. Care was taken to avoid mis-assigning alleles when combining results from different samples and platforms.37 Using the ‘STRANDLABEL’ and ‘USESTRAND ON’ commands in METAL, these SNPs were recoded to ensure consistent allele coding across the samples analyzed. Because the German sample was genotyped on the Illumina platform (San Diego, CA, USA) that contains no palindromic SNPs, we used that sample as the gold standard for our study.

Results were combined under a fixed-effects model, using METAL. For initial discovery purposes, the fixed-effects model is more powerful than the traditional random-effects model, and Pereira et al.24 suggest the fixed-effects model is preferable, especially when the cumulative sample size is in the range of 2000–20 000.

After selected results were confirmed, heterogeneity statistics were calculated, using Comprehensive Meta-analysis version 2.0. When heterogeneity tests are significant, assumptions of the fixed-effects model are violated, and the results can be anti-conservative.24 To address this, we also analyzed selected SNPs under a random-effects model. As the traditional random-effects method assumes a markedly conservative null-hypothesis model,38 we used a novel method, implemented in Metasoft (vers3.1),39 designed to increase power when there is true heterogeneity.

We used a threshold of genome-wide significance of P=5 × 10−8 derived from a published, genome-wide simulation of common variants in samples of European ancestry.40, 41, 42 Marginally significant (P<1 × 10−6) SNPs are also reported. Random-effects P-values were used when the Q-test of heterogeneity was significant at P<0.1.43 Although we had full access to complete genome-wide results for all phase I samples, we had access to the GSK and FondaMental results for only the most significant SNPs identified in phase I. As joint analysis has been shown to be more powerful than replication analysis in situations like this,44 the most significant markers from phase I (P<4 × 10−3) were combined with the additional phase II samples for the joint analysis.

Power analysis

Power Analysis was done with genetic power calculator.45 We assumed a trait prevalence of 2%, minor allele frequency of 0.2, alpha of 5 × 10−8 and a marker-allele D′ value of 0.9. The phase I sample had >80% power to detect an allele that confers a genotype-relative risk of 1.25 under a log-additive model.

Drug treatment of cell cultures and quantitative real-time PCR (qPCR)

HeLa, SH-SY5Y and HEK293 cells were grown in dulbecco's modified Eagles's medium with 4.5 gm l–1 d-glucose supplemented with 2 mM L-glutamine and 10% fetal calf serum. Lithium, valproic acid (VPA), dexamethasone and triamcinolone were purchased from Sigma-Aldrich (St Louis, MO, USA). Drug treatments were conducted in HeLa, SH SY5Y and HEK293 cells.46, 47, 48, 49 Total RNA was extracted using RNeasy plus Mini-kits (QIAGEN, Valencia, CA, USA). First-strand cDNA was synthesized with First Strand Superscript III kits (Invitrogen, Carlsbad, CA, USA). mRNA levels of selected genes were determined by using Roche LightCycler 480 (F.Hoffmann-La Roche Ltd, Basel, Switzerland) and the Roche Universal Probe Library System.

To help in clarifying the genetic role of TRANK1, we examined gene expression in cells after drug treatment.47, 48 SH-SY5Y cells were treated with lithium carbonate (1 or 2 mM), VPA (0.5 or 5 mM), dexamethasone (200 nM or 2 μM) or triamcinolone (3,10, or 100 μM) for 24 h. VPA treatment was also given at 0.5 mM or 5 mM for 75 h. Total RNA was isolated and quantitative real-time PCR (qPCR) was performed using LightCycler 480 (Roche). The ΔΔCT method was used to quantify relative mRNA levels. Any differences were tested with the Student's t-test. Data represent mean±s.e.m. from three independent experiments.

Analysis of eQTL from human postmortem brain sample data sets

Postmortem brain tissues from two collections,50, 51, 52 totaling 243 samples, were used to look at the potential role of the top-hit SNPs in nearby gene regulation. Gibbs et al.50 and Heinzen et al.51 describe these sample –data sets in detail. Gibbs et al. provide both the uncorrected and empirical P-value. For Heinzen et al. in SNPExpress, significance threshold was calculated on the basis of the total number of association tests conducted within the analyses. The top hits were examined for cis-effects in SNPExpress, and in Gibbs et al. Bonferroni correction was applied for the total number of top hits tested.


Overall, 748 555 SNP markers were consistently scored across the six phase I samples. The genome-wide mean Z-score was <0.004, approximating the theoretical mean of zero. This indicates an unbiased experiment.

Analysis of the combined European and Asian ancestry phase I sample detected genome-wide significant evidence of association between BD and three SNPs located near two different genes (Figures 1 and 2). At rs9834970, the C-allele was consistently more common in cases than controls (P=2.41 × 10−11 Table 2a; Supplementary Tables 2, and 3A). Several nearby SNPs in linkage disequilibrium with rs9834970 were significant at the P<4 × 10−3 level (Figure 2). These SNPs are clustered within and around the gene TRANK1 (LBA1) on chromosome 3p22.2.

Figure 1

Meta-analysis flow-diagram. Described on the left is the world-wide discovery collection consisting of European plus Asian-ancestry meta-analysis, both at the phase I sample analysis level as well at the phase I sample and phase II sample joint analysis level. On the right of the diagram, a similar European-ancestry only meta-analysis, is likewise described. At both the level of phase I analysis and phase I sample and phase II sample joint analysis, genome-wide significant (P<5 × 10−8) SNPs are described as well as those which were found to be marginally significant (P<1 × 10−6; for detailed list, lease refer to Table 2 and Supplementary Tables 2 and 3). For easy reference, names of nearest genes are given.

PowerPoint slide

Figure 2

(a) Manhattan plot of the European plus Asian-ancestry phase I sample meta-analysis results, generated by Haploview 4.2. Physical position is shown along the X-axis with each chromosome shown in distinct color; −log (meta-P-value) is shown along the Y-axis. The red guideline indicates the threshold of genome-wide significance (5 × 10−8). The blue line indicates suggestive (P<1 × 10−6) results. (b) Detail of the associated region for rs2271893, generated by SNAP 2.2. Physical position and gene annotations (1000 Genome Pilot I) are shown along the X-axis, −log (meta-P value) is shown on the left Y-axis, recombination rate (CEU) on the right Y-axis; Below, linkage disequilibrium (r2) as estimated from HapMap 3 phased genotypes, generated by UCSC Genome Browser. Darker red indicates higher values. Recombination rates (CHB/JPT) are not significantly different. (c) Similarly, detail of the associated region for rs9834970, generated by SNAP 2.2. Physical position and gene annotations (1000 Genome Pilot I) are shown along the X-axis, −log (meta-P value) is shown on the left Y-axis, recombination rate (CEU) on the right Y-axis; Below, linkage disequilibrium (r2) as estimated from HapMap 3-phased genotypes, generated by UCSC Genome Browser. Darker red indicates higher values. Again, recombination rates (CHB/JPT) are not significantly different.

PowerPoint slide

Table 2 Meta-analysis results by ancestrya

Two additional SNPs, rs2271893 and rs6746896, also returned genome-wide significant evidence of association with BD (P<4 × 10−8). Several nearby SNPs in linkage disequilibrium with rs2271893 and rs6746896 were significant at the P<4 × 10−3 level. These SNPs are located on chromosome 2q11.2 in the vicinity of LMAN2L (Figure 2).

An additional 36 SNPs were significant at the P<10−6 level. These include SNPs near ANK3 and the 3p21.1 locus, reported in previous studies.12, 13, 17 These and the three genome-wide significant SNPs were all taken forward into the phase II analysis.

After addition of the two additional samples in the phase II meta-analysis, the overall picture was similar. Genome-wide significant association signals remained at rs9834970 (random effects P=1.48 × 10−12) and rs2271893 (random effects P=2.2 × 10−10), although the heterogeneity test revealed significant results (Table 2b). At these two SNPs, the direction of the association signal was consistent in the FondaMental sample, but reversed (non-significantly) in the GSK sample (Supplementary Tables 2, and 3B). Additional genome-wide significant signals were detected at rs4650608 near PTGFR; at rs4948418, near ANK3; and at rs7618915, within the 3p21 locus.

Analysis of only the European-ancestry samples gave similar results. SNP rs9834970 near TRANK1 remained genome-wide significant (P=5.65 × 10−10), along with rs2271893 and rs6746896 near LMAN2L (P=1.21 × 10−9, 1.66 × 10−9, respectively). Several regions contained SNPs with marginally significant associations (5 × 10−8<P<1 × 10−6), including a set of SNPs in linkage disequilibrium at 3p21.1 (Figure 1; Table 2c; Supplementary Table 3C). The phase II analysis similarly identified the same genome-wide significant SNPs near TRANK1, LMAN2L and PTGFR (Figure 1; Table 2d; Supplementary Table 3D).

Heterogeneity analysis was significant (P<0.1) for several SNPs (Table 2; Supplementary Figure 1). The Asian-ancestry sample did not significantly contribute to heterogeneity as measured by I2. We observed no significant heterogeneity for rs9834970 within the phase I samples (Q=3.95, P=0.267; I2=23.98; Table 2). However, the heterogeneity test became significant in the phase II analysis (Q=10.95, P=0.052; I2=54.34; Table 2), owing to the GSK sample.

We explored the possible role of the top-hit SNPs in regulating expression of nearby genes (Supplementary Table 5).50, 51, 52 SNP rs9834970 was not available in either of the brain-expression data sets we queried. The SNPs rs2251219, rs4650608 and rs6746896 were found to show significance after Bonferroni correction in one data set.51 None of the top hits nor their proxies with r2>0.4 were represented in Gibbs et al.50 Interestingly, rs4650608 was not associated with expression of the closest gene, PTGFR, but rather with IFI44L, an interferon-induced protein with possible increased expression in the hypothalamus.53

SNP rs9834970 is located about 12 kb distal to the 3′ UTR of TRANK1 (hg18), recently known as LBA1. To explore a possible functional role of TRANK1 in BD, we assessed TRANK1 expression after exposing cells to established treatments for BD. TRANK1 expression was measured in vitro after treatment with the mood stabilizers lithium and VPA, as well as the corticosteroids dexamethasone and triamcinolone, sometimes known to provoke manic episodes in BD patients.54 We tested TRANK1 mRNA expression in three separate cell lines. As shown in Figure 3 and Supplementary Figure 2, VPA markedly increased TRANK1 mRNA expression in a dose- and time-dependent manner, increasing expression from very low baseline levels. Lithium, dexamethasone, and triamcinolone had no measurable effect on the low levels of TRANK1 expression in the cell lines we tested.

Figure 3

Dose-and time-dependent changes in TRANK1 expression after valproic acid (VPA) treatment in SHSY5Y (a), HeLa (b) and HEK293 (c) cells. Significance was determined by using Student's t test. Data are mean±s.e.m. from three independent experiments. *P<0.05; **P<0.01 compared with control. As shown, VPA increased TRANK1 mRNA expression in a dose- and time-dependent manner in the three cell lines.

PowerPoint slide

We also queried this data set for previously published genome-wide significant BD markers (Supplementary Table 4).12, 13, 17, 55, 56, 57 All published, BD-associated markers were available in our meta-analysis except rs10994336. rs2251219 was significant at P<1.84 × 10−7 level. None of the other SNPs were significant beyond the P<3 × 10−4.


We describe, in a worldwide collection, genome-wide significant evidence in support of three novel genetic markers associated with BD. The results suggest novel association findings near the genes TRANK1 (LBA1), LMAN2L and PTGFR. We show that VPA markedly increased TRANK1 mRNA expression in vitro in a dose- and time-dependent manner, supporting a functional role for TRANK1 in BD treatment. We also find evidence consistent with previous studies, supporting association between BD and SNPs at a chromosome 3p21 locus and near ANK3. These results are also broadly consistent with those based on other large samples (Cichon et al.;55 Psychiatric GWAS Consortium,58), although individual markers in each study vary in significance between the P<10−5 and P<10−9 levels. As discussed below, much larger sample sizes may be needed before consistent genome-wide significance is attained for any one marker.

How many more markers like those already found might eventually be identified by GWAS methods? We used the method by Park et al.59 to estimate the number of underlying susceptibility loci that are likely to reside within the range of effect sizes observed in the current and in previous GWAS.59 We selected the five novel SNPs that showed association with BD at P<5 × 10−7 in the combined phase I European and Asian ancestry samples. A bias-correction method, described by Ghosh et al.,60 was applied to the corresponding odds ratio estimates to address the ‘Winner's Curse’.60 We also included five independent, previously-identified SNPs (or proxies, r2>0.8) with association P-values <4 × 10−3 in phase I. As the previously identified SNPs showed genome-wide significant associations in other studies, odds ratio estimates for these SNPs were taken directly from the phase I results without correction, under the assumption that phase I serves as a replication sample. Once power for detection was calculated based on the design of the current study, the number of underlying loci—and the trajectory for their discovery as sample sizes grow—was estimated by an inverse-power weighted approach.59

The BD discovery trajectory has several interesting properties (Figure 4): (1) Assuming equal numbers of case–controls, we estimated that a total sample size of 16 000 is needed to discover 10 independent-susceptibility loci at the P<5 × 10−7 level (Table 2b; Supplementary Tables 3B and 6); (2) We observed that 10 susceptibility loci account for 1% of total genetic variance; (3) About 19 000 additional individuals may be needed to discover markers explaining the next 2% of genetic variance; (4) The identification of SNPs that account for the total 5.5% of genetic variance would require at least 63 000 total individuals; (5) The sample size required to discover half the total number of susceptibility loci—a value we designate the ND50—is about 35 000–45 000. We would like to point out that our confidence in the discovery trajectory is limited by the fact that only 10 observed BD susceptibility loci could be employed in estimation of the model; the new discovery trajectory will become more reliable as additional susceptibility loci are identified.59

Figure 4

Estimated novel locus discovery (ND) trajectory for BD with additional individuals (AI) collected. The filled marker indicates present level of observed data. Cumulative expected loci estimate is depicted on the X-axis. The left Y-axis describes the percentage of genetic variance accounted for by these cumulative expected loci. The right Y-axis shows estimated sample size required for the expected number of loci and variance explained. (1) With 16 000 total individuals (TI), assuming equal case–controls, 10ND is observed; (2) >63 000 TI is estimated for discovery of the cumulative expected number of loci.

PowerPoint slide

The total sample size used in this meta-analysis, although large, still limited the power to detect novel risk loci. Previous studies of meta-analysis for discovery purposes in common traits have approximated 66 000 for case–control studies of low-prevalence/high-heritability diseases, compared with a 2.4-fold-increase in samples required for high-prevalence/low-heritability diseases.12, 61, 62, 63 For analysis of the population's quantitative traits, sample size would need to exceed 120 000 to achieve equivalent discovery power. Our own estimates describe similar increases in discovery of novel loci as sample sizes grow (Figure 4).

On the basis of the cumulative GWAS findings to date, we estimate that a 2-fold-increase in total sample size will be needed to reach the ND50 for BD—a challenging task, but not impossible. We estimate a minimum of 105 common BD susceptibility loci could be found by GWAS. With the identification of uncommon and rare variants, for example, by large-scale sequencing, the total genetic variance explained should increase, along with the number of risk loci.59

Findings in the combined European and Asian-ancestry samples largely agreed with those in the European-only analysis. Such concordant findings, spanning continental populations, may suggest shared risk alleles.64 In 1904, Emil Kraepelin conducted a comparative psychiatric study in the asylum of Buitenzorg, in Java, most of whose indigenous people were then relatively isolated from European influences. Kraepelin65 wrote that he saw ‘…no compelling grounds for assuming that the natives of Java suffer from new and unknown forms of insanity.’ In 2010, Pritchard et al.,64 using HapMap data to estimate the influence of natural selection on genetic adaptation, observed few fixed, evolutionarily-selected alleles, instead finding steady, gradual variation in allele frequencies across continents. This gradual divergence in allele frequencies between people of West-Eurasian and East-Asian origins may have occurred as recently as 60 000 YBP. Our results are broadly consistent with these ideas.

We did detect significant heterogeneity, although this appeared to be driven in this study by the GSK sample. Power is low with such small samples, and it is difficult to draw strong conclusions concerning heterogeneity. It may be important that the GSK sample is largely non-familial, whereas most of the other samples consist of probands collected with family-based ascertainment. Further studies are needed to understand all the sources of heterogeneity that may contribute here.

It has been typical in the meta-analysis literature to restrict fixed-effects models to situations where the heterogeneity test is not significant. However, Pereira et al.24 have shown that random-effects models are much less powerful when the total sample size is below 20 000. They conclude that ‘… fixed effects may be preferable for the purposes of initial discovery, if the aim is simply to screen and identify as many of the true variants as possible,’ consistent with the objective of our study. They further note that ‘…with fixed-effects, the rate of false positives increases substantially as more data accumulate’ and recommend that ‘associations that pass desired significance thresholds with fixed-effects calculations, but not with random-effects calculations, may require further replication.’ For this reason, we used a random-effects test whenever the heterogeneity test was significant. Further testing in large samples will be needed before we can claim replication.

The Wellcome Trust Case–Control Consortium previously suggested association with BD for rs9834970 (P=4.50 × 10−7). It is interesting to note that association between BD and rs984790 was recently reported at experiment-wide significance in an additional, independent sample collected in Brazil.66 We did not include this sample in our main meta-analysis as genome-wide data were not available. The effect size was quite large in this small family sample (odds ratio=2.64), and the direction of association was consistent with our findings. When the Brazilian data were combined with the European and Asian-ancestry samples, the overall significance increased further, along with the evidence for heterogeneity of effect sizes (fixed-effects P=1.54 × 10−14, random-effects P=1.61 × 10−14, Q=16.47, P=0.011, I2=63.58; Supplementary Table 1). This illustrates an interesting property of the heterogeneity tests commonly employed in meta-analysis: Heterogeneity can increase when a sample shows a much larger effect size than other samples in the analysis.

Taken together, these data support TRANK1 as a novel risk locus for BD. TRANK1, formerly named, KIAA0342 and LBA1, stands for tetratricopeptide repeat and ankyrin repeat containing 1. It is a roughly 34 kb (14 total exon) gene on the reverse strand of chromosome 3p22.2 that encodes a P-loop containing nucleoside triphosphate hydrolase, associated with DNA/ATP binding or DNA repair, with significant expression in the brain.53 The dose and time-dependent increase in TRANK1 mRNA expression that we observed in response to VPA treatment in three separate cell lines provides an independent line of supportive evidence, suggesting that the causal variant(s) in TRANK1 lead to a loss of function. Future expression studies of TRANK1 as well as sequencing may be fruitful future directions.

Genome-wide significant signals were also observed for SNPs near LMAN2L, PTGFR, the 3p21 locus and ANK3, but with significant evidence of heterogeneity in the final analysis. LMAN2L, also known as lectin, mannose-binding 2-like, located on chromosome 2, is hypothesized to be involved in the regulation of export from the endoplasmic reticulum of a subset of glycoproteins, and may also serve as a regulator of ERGIC-53.53, 67 PTGFR on chromosome 1p31.1 encodes the prostaglandin F receptor, a member for the G-protein-coupled receptor family.67, 68 Both LMAN2L and PTGFR are highly expressed in the brain.53, 67 Further replication and expression studies will be needed to validate these findings.

These results support a few of the previous GWAS findings and add three new candidate genes to a growing list of BD risk loci. It will be necessary to increase the set of robustly identified risk loci further if we are to succeed in using these findings to triangulate the biochemical pathways involved in the etiology of BD. Sample size will remain a limiting factor in the discovery of common alleles associated with BD by GWAS. Methods that interrogate rarer alleles, such as next-generation sequencing, could help fill the remaining heritability gap, as suggested by recent studies.69 Clearly, there are many BD risk loci yet to discover. As they accumulate, they will shed light on the biology of BD and help lead us toward better treatments.

The full set of results is available for download at


  1. 1

    Bertelsen A, Harvald B, Hauge M . A Danish twin study of manic-depressive disorders. Br J Psychiatry 1977; 130: 330–351.

  2. 2

    Kendler KS, Pedersen N, Johnson L, Neale MC, Mathe AA . A pilot Swedish twin study of affective illness, including hospitaland population-ascertained subsamples. Arch Gen Psychiatry 1993; 50: 699–700.

  3. 3

    Kendler KS . Twin studies of psychiatric illness: an update. Arch Gen Psychiatry 2001; 58: 1005–1014.

  4. 4

    Cardno AG, Marshall EJ, Coid B, Macdonald AM, Ribchester TR, Davies NJ et al. Heritability estimates for psychotic disorders: the Maudsley Twin Psychosis Series. Arch Gen Psychiatry 1999; 56: 162–168.

  5. 5

    Edvardsen J, Torgersen S, Roysamb E, Lygren S, Skre I, Onstad S . Heritability of bipolar spectrum disorders. Unity or heterogeneity? J Affect Disord 2008; 106: 229–240.

  6. 6

    Smoller JW, Finn CT . Family, twin, and adoption studies of bipolar disorder. Am J Med Genet C Semin Med Genet 2003; 123C: 48–58.

  7. 7

    Wendland JR, McMahon FJ . Genetics of Bipolar Disorder. In: Manji HK and Zarate Jr CA (eds). Behavioral Neurobiology of Bipolar Disorder and its Treatment. Springer: London, UK, 2011 pp 19–31.

  8. 8

    Gershon ES, Hamovit J, Guroff JJ, Dibble E, Leckman JF, Sceery W et al. A family study of schizoaffective, bipolar I, bipolar II, unipolar and normal control probands. Arch Gen Psychiatry 1982; 39: 1157–1167.

  9. 9

    Kieseppä T, Partonen T, Haukka J, Kaprio J, Lönnqvist J . High concordance of bipolar I disorder in a nationwide sample of twins. Am J Psychiatry 2004; 161: 1814–1821.

  10. 10

    Goodwin FK, Jamison K . Manic-Depressive Illness: Bipolar Disorders and Recurrent Depression, 2nd Edn. Oxford Univ Press: New York, USA, 2007.

  11. 11

    Craddock N, Jones I . Genetics of bipolar disorder. J Med Genet 1999; 36: 585–594.

  12. 12

    Ferreira MA, O'Donovan MC, Meng YA, Jones IR, Ruderfer DM et al. Collaborative genome-wide association analysis supports a role for ANK3 and CACNA1C in bipolar disorder. Nature Genet 2008; 40: 1056–1058.

  13. 13

    Baum AE, Akula N, Cabanero M, Cardona I, Corona W, Klemens B et al. A genome-wide association study implicates diacylglycerol kinase eta (DGKH) and several other genes in the etiology of bipolar disorder. Mol Psychiatry 2007; 13: 197–207.

  14. 14

    Schulze TG, Detera-Wadleigh SD, Akula N, Gupta A, Kassem L, Steele J et al. Two variants in Ankyrin 3 (ANK3) are independent genetic risk factors for bipolar disorder. Mol Psychiatry 2009; 14: 487–491.

  15. 15

    Scott LJ, Muglia P, Kong XQ, Guan W, Flickinger M, Upmanyu R et al. Genome-wide association and meta-analysis of bipolar disorder in individuals of European ancestry. Proc Natl Acad Sci USA 2009; 106: 7501–7506.

  16. 16

    Sklar P, Smoller JW, Fan J, Ferreira MA, Perlis RH, Chambert K et al. Whole-genome association study of bipolar disorder. Mol Psychiatry 2008; 13: 558–569.

  17. 17

    McMahon FJ, Akula N, Schulze TG, Pierandrea M, Tozzi F, Detera-Wadleigh SD et al. Meta-analysis of genome-wide association data detects a risk locus for major mood disorders on chromosome 3p21.1. Nat Genet 2010; 42: 128–131.

  18. 18

    Zeggini E, Scott LJ, Saxena R, Voight BF, Marchini JL, Hu T et al. Meta-analysis of genome-wide association data and large-scale replication identifies additional susceptibility loci for type 2 diabetes. Nat Genet 2008; 40: 638–645.

  19. 19

    Zeggini E, Weedon MN, Lindgren CM, Frayling TM, Elliott KS, Lango H et al. Replication of genome-wide association signals in UK samples reveals risk loci for type 2 diabetes. Science 2007; 316: 1336–1341.

  20. 20

    Weedon MN, Lango H, Lindgren CM, Wallace C, Evans DM, Mangino M et al. Genome-wide association analysis identifies 20 loci that influence adult height. Nat Genet 2008; 40: 575–583.

  21. 21

    Willer CJ, Sanna S, Jackson AU, Scuteri A, Bonnycastle LL, Clarke R et al. Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet 2008; 40: 161–169.

  22. 22

    Houlston RS, Webb E, Broderick P, Pittman AM, Di Bernardo MC, Lubbe S et al. Meta-analysis of genome-wide association data identifies four new susceptibility loci for colorectal cancer. Nat Genet 2008; 40: 1426–1435.

  23. 23

    Raychaudhuri S, Remmers EF, Lee AT, Hackett R, Guiducci C, Burtt NP et al. Common variants at CD40 and other loci confer risk of rheumatoid arthritis. Nat Genet 2008; 40: 1216–1223.

  24. 24

    Pereira TV, Patsopoulos NA, Salanti G, Ioannidis JP . Discovery properties of genome-wide association signals from cumulatively combined data sets. Am J Epidemiol 2009; 170: 1197–1206.

  25. 25

    Moreno C, Laje G, Blanco C, Jiang H, Schmidt AB, Olfson M . National trends in the outpatient diagnosis and treatment of bipolar disorder in youth. Arch Gen Psychiatry 2007; 64: 1032–1039.

  26. 26

    Pritchard JK, Coop G, Pickrell JK, Novembre J, Kudaravalli S, Li J et al. The role of geography in human adaptation. PLoS Genet 2009; 5: 1–16.

  27. 27

    de Bakker PI, Neale BM, Daly MJ . Meta-analysis of genome-wide association studies. Cold Spring Harb Protoc 2010; doi:10.1101/pdb.top81;

  28. 28

    de Bakker PI, Ferreira MA, Jia X, Neale BM, Raychaudhuri S, Voight BF . Practical aspects of imputation-driven meta-analysis of genome-wide association studies. Hum Mol Genet 2008; 17: R122–R128.

  29. 29

    Smith EN, Bloss CS, Badner JA, Barrett T, Belmonte PL, Berrettini W et al. Genome-wide association study of bipolar disorder in European American and African American individuals. Mol Psychiatry 2009; 14: 755–763.

  30. 30

    Lee MTM, Chen CH, Lee CH, Chen CC, Chong MY, Ouyang WC et al. Genome-wide association study of bipolar I disorder in Han Chinese Population. Mol Psychiatry 2010; 16: 548–556.

  31. 31

    Etain B, Dumaine A, Mathieu F, Chevalier F, Henry C, Kahn JP et al. A SNAP25 promoter variant is associated with early-onset bipolar disorder and a high expression level in brain. Mol Psychiatry 2010; 15: 748–755.

  32. 32

    Sebastiani P, Solovieff N, Puca A, Hartley SW, Melista E, Andersen S et al. Genetic signatures of exceptional longevity in humans. Science 2010; e-pub ahead of print 11 November 2010; doi:10.1126/science.1190532.

  33. 33

    Li Y, Abecasis GR . MACH 1.0: Rapid haplotype reconstruction and missing genotype inference. Am J Hum Genet 2006; S79: 2290.

  34. 34

    Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 559–575.

  35. 35

    Browning BL, Browning SR . A unified approach to genotype imputation and haplotype phase inference for large data sets of trios and unrelated individuals. Am J Hum Genet 2008; 84: 210–223.

  36. 36

    Devlin B, Roeder K . Genomic control for association studies. Biometrics 1999; 55: 997–1004.

  37. 37

    Ustun TB, Ayuso-Mateos JL, Chatterji S, Mathers C, Murray CJ . Global burden of depressive disorders in the year 2000. Br J Psychiatry 2004; 184: 386–392.

  38. 38

    Han B, Eskin E . Random-effects model aimed at discovering associations in meta-analysis of genome-wide association studies. Am J Hum Genet 2011; 88: 586–598.

  39. 39

    Metasoft vo3.1 [Internet]. [cited 2011 July 7] [Place Unknown] UCLA, Available from: 2011.

  40. 40

    Pe'er I, Yelensky R, Altshuler D, Daly MJ . Estimation of the multiple testing burden for genome-wide association studies of nearly all common variants. Genet Epidemiol 2008; 32: 381–385.

  41. 41

    Dudbridge F, Gusnanto A . Estimation of significance thresholds for genome-wide association scans. Genet Epidemiol 2008; 32: 227–234.

  42. 42

    Manolio TA, Brooks LD, Collins FS . A HapMap harvest of insights into the genetics of common disease. J Clin Invest 2008; 118: 1590–1605.

  43. 43

    Ioannidis JP . Interpretation of tests of heterogeneity and bias in meta-analysis. J Eval Clin Pract 2008; 14: 951–957.

  44. 44

    Skol AD, Scott LJ, Abecasis GR, Boehnke M . Joint analysis is more efficient than replication-based analysis for two-stage genome-wide association studies. Nat Genet 2006; 38: 209–213.

  45. 45

    Purcell S, Cherny SS, Sham PC . Genetic Power Calculator: design of linkage and association genetic mapping studies of complex traits. Bioinformatics 2003; 19: 149–150.

  46. 46

    Phiel CJ, Zhang F, Huang EY, Guenther MG, Lazar MA, Klein PS et al. Histone deacetylase is a direct target of valproic acid, a potent anticonvulsant, mood stabilizer, and teratogen. J Biol Chem 2001; 276: 36734–36741.

  47. 47

    Schüle C, Baghai TC, Eser D, Häfner S, Born C, Herrmann S et al. 2009) The combined dexamethasone/CRH test (DEX/CRH test) and prediction of acute treatment response in major depression. PLoS ONE 2009; 4: e4324.

  48. 48

    Barbier E, Wang JB . Anti-depressant and anxiolytic like behaviors in PKCI/HINT1 knockout mice associated with elevated plasma corticosterone level. BMC Neurosci 2009; 10: 132.

  49. 49

    Gilmor ML, Skelton KH, Nemeroff CB, Owens MJ . The effects of chronic treatment with the mood stabilizers valproic acid and lithium on corticotropin-releasing factor neuronal systems. J Pharmacol Exp Ther 2003; 305: 434–439.

  50. 50

    Gibbs JR, van der Brug MP, Hernandez DG, Travnor BJ, Nalls MA, Lai SL et al. Abundant quantitative trait loci exist for DNA methylation and gene expression in human brain. PLoS Genet 2010; 6: e1000952.

  51. 51

    Heinzen E, Ge D, Cronin KD, Maia J, Shianna KV, Gabriel W et al. Tissue specific genetic control of gene expression and alternative splicing: Implications for the study of human complex traits. PLoS Biol 2008; 6: e1.

  52. 52

    SNPExpress [Internet]. Duke University 2011. Available from:

  53. 53

    Kent WJ, Sugnet CW, Furey TS, Roskin KM, Pringle TH, Zahler AM et al. The human genome browser at UCSC. Genome Res 2002; 12: 996–1006.

  54. 54

    David AS, Fleminger S, Kopelman MD, Lovestone S, Mellers JDC (eds). Lishman's Organic Psychiatry: A Textbook of Neuropsychiatry, 4th edn. Wiley-Blackwell: Chichester, 2009.

  55. 55

    Cichon S, Mühleisen TW, Degenhardt FA, Mattheisen M, Miró X, Strohmaier J et al. Genome-wide association study identifies genetic variation in neurocan as a susceptibility factor for bipolar disorder. Am J Hum Genet 2011; 88: 372–381.

  56. 56

    Wellcome Trust Case Control Consortium. Genome-wide association study of 14 000 cases of seven common diseases and 3000 shared controls. Nature 2007; 447: 661–678.

  57. 57

    O'Donovan MC, Craddock N, Norton N, Williams H, Peirc T, Moskvina V et al. Identification of loci associated with schizophrenia by genome-wide association and follow-up. Nat Genet 2008; 40: 1053–1055.

  58. 58

    Psychiatric GWAS Consortium Bipolar Working Group, Sklar P, Ripke S, Scott L, Andreassen OA, Cichon S et al. Large-scale genome-wide association analysis of bipolar disorder identifies a new susceptibility locus near ODZ4. Nat Genet 2011; 43: 977–983.

  59. 59

    Park JH, Wacholder S, Gail MH, Peters U, Jacobs KB, Chanock SJ et al. Estimation of effect size distribution from genome-wide association studies and implications for future discoveries. Nature Genet 2010; 42: 570–575.

  60. 60

    Ghosh A, Zou F, Wright FA . Estimating odds ratios in genome scans: an approximate conditional likelihood approach. Am J Hum Genet 2008; 82: 1064–1074.

  61. 61

    Wray NR, Pergadia ML, Blackwood DHR, Penninx BWJH, Gordon SD, Nyholt DR et al. Genome-wide association study of major depressive disorder: New results, meta-analysis, and lessons learned. Mol Psychiatry 2010; e-pub ahead of print 2 November 2010. doi: 10.1038/mp.2010.109.

  62. 62

    Green EK, Grozeva D, Jones I, Jones L, Kirov G, Caesar S et al. The bipolar disorder risk allele at CACNA1C also confers risk of recurrent major depression and of schizophrenia. Mol Psychiatry 2010; 15: 1016–1022.

  63. 63

    Yang J, Wray NR, Visscher PM . Comparing apples and oranges: equating the power of case–control and quantitative trait association studies. Genet Epidemiol 2010; 34: 254–257.

  64. 64

    Pritchard JK, Pickrell JK, Coop G . The genetics of human adaptation: hard sweeps, soft sweeps, and polygenic adaptation. Curr Biol 2010; 20: R208–R215.

  65. 65

    Kraeplin E . Comparative Psychiatry. In: Hirsch SR (eds). Themes and Variations in European Psychiatry. The University Press of Virginia: Charlottesville, pp 3–30.

  66. 66

    Secolin R, Banzato CEM, Oliveira MCM, Bittar MFR, Santos ML, Dalgalarrondo P et al. Family-based association study for bipolar affective disorder. Psychiatr Genet 2010; 20: 126–129.

  67. 67

    Yanai I, Benjamin H, Shmoish M, Chalifa-Caspi V, Shklar M, Ophir R et al. Genome-wide midrange transcription profiles reveal expression level relationships in human tissue specification. Bioinformatics 2005; 5: 650–659.

  68. 68

    GeneCards [Internet]. [cited 11 May 2011] Weizmann Institute of Science 2011. Available from:

  69. 69

    Lee SH, Wray NR, Goddard ME, Visscher PM . Estimating missing heritability for disease from genome-wide association studies. Am J Hum Genet 2011; 88: 294–305.

  70. 70

    Littell JH, Corcoran J, Pillai VK . Systematic Reviews and Meta-Analysis. Oxford University Press: New York, USA, 2008.

Download references

Author information

Correspondence to D T Chen.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Molecular Psychiatry website

PowerPoint slides

Supplementary information



The Bipolar Genome Study (BiGS) Authorship List

University California, San Diego: John R. Kelsoe, Tiffany A. Greenwood, Caroline M. Nievergelt, Rebecca McKinney, Paul D. Shilling; Scripps Translational Science Institute: Nicholas J. Schork, Erin N. Smith, Cinnamon S. Bloss; Indiana University: John I. Nurnberger, Jr., Howard J. Edenberg, Tatiana Foroud, Daniel L. Koller; University of Chicago: Elliot S. Gershon, Chunyu Liu, Judith A. Badner; Rush University Medical Center: William A. Scheftner; Howard University: William B. Lawson, Evaristus A. Nwulia, Maria Hipolito; University of Iowa: William Coryell; Washington University: John Rice; University California, San Francisco: William Byerley; National Institute Mental Health: Francis J. McMahon, David T Chen; University of Pennsylvania: Wade H. Berrettini; Johns Hopkins: James B. Potash, Peter P. Zandi, Pamela B. Mahon; University of Michigan: Melvin G. McInnis, Sebastian Zöllner, Peng Zhang; The Translational Genomics Research Institute: David W. Craig, Szabolcs Szelinger; Portland Veterans Affairs Medical Center: Thomas B. Barrett; Georg-August-University Göttingen: Thomas G. Schulze.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Chen, D., Jiang, X., Akula, N. et al. Genome-wide association study meta-analysis of European and Asian-ancestry samples identifies three novel loci associated with bipolar disorder. Mol Psychiatry 18, 195–205 (2013).

Download citation


  • ANK3
  • bipolar disorder
  • LBA1
  • meta-analysis
  • TRANK1
  • 3p21

Further reading