Schizophrenia (SZ) is a neurodevelopmental disorder in which altered immune function typically plays an important role in mediating the effect of environmental insults and regulation of inflammation. The breast cancer suppressor protein associated protein (BRAP) is suggested to exert vital effects in neurodevelopment by modulating the mitogen-activated protein kinase cascade and inflammation signaling. To explore the possible role of BRAP in SZ, we conducted a two-stage study to examine the association of BRAP polymorphisms with SZ in the Han Chinese population. In stage one, we screened SNPs in BRAP from our GWAS data, which detected three associated SNPs, with rs3782886 being the most significant one (P = 2.31E-6, OR = 0.67). In stage two, we validated these three SNPs in an independently collected population including 1957 patients and 1509 controls, supporting the association of rs3782886 with SZ (P = 1.43E-6, OR = 0.73). Furthermore, cis-eQTL analysis indicates that rs3782886 genotypes are associated with mRNA levels of aldehyde dehydrogenase 2 family (ALDH2) (P = 0.0039) and myosin regulatory light chain 2 (MYL2) (P < 1.0E-4). Our data suggest that the BRAP gene may confer vulnerability for SZ in Han Chinese population, adding further evidence for the involvement of developmental and/or neuroinflammatory cascades in the illness.
Citation: Zhang F, Liu C, Xu Y, Qi G, Yuan G, Cheng Z, et al. (2014) A Two-Stage Association Study Suggests BRAP as a Susceptibility Gene for Schizophrenia. PLoS ONE 9(1): e86037. https://doi.org/10.1371/journal.pone.0086037
Editor: Chunyu Liu, University of Illinois at Chicago, United States of America
Received: August 29, 2013; Accepted: December 4, 2013; Published: January 15, 2014
Copyright: © 2014 Zhang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by research grants from the National High-Tech Research and Development Program of China (2009AA022702), the National Natural Science Foundation of China (81000578, 81071087 and 81071088), Science Fund for Creative Research Groups of the National Natural Science Foundation of China (81221002), the National Basic Research Program of China (2011CB707805), and the International Science & Technology Cooperation Program of China (2010DFB30820). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Schizophrenia (SZ) is a common yet disabling mental disorder characterized by profound disruption in cognition and emotion, affecting the most fundamental human attributes: language, thought, perception, affect, and sense of self. The onset of the illness is typically in late adolescence or early adulthood. Epidemiological research show that a variety of pre- and peri-natal insults are associated with increased vulnerability to SZ .
Since its initial proposition in late 1980’s , , the neurodevelopmental hypothesis has received much support from epidemiological, developmental and neuroimaging studies  and has been the dominant paradigm for schizophrenia research over the past two decades . This hypothesis posits that SZ has its roots in disturbed development of the nervous system, in which cerebral insults occur during early brain development long before the full-blown of the illness. On the other hand, immune system may modulate normal neurodevelopment and there is increasing evidence for altered inflammatory factors in the etiology and pathophysiology of SZ . For example, SZ has been shown to be associated with activated peripheral and central inflammatory responses; cytokines such as interleukin (IL)-1β, IL-6, and tumor necrosis factor α (TNF-α) participate in regulating normal brain development and have been implicated in abnormal corticogenesis .
The developmental formation of the mammalian central nervous system (CNS) requires precise spatial and temporal integration of mitogenic and neurogenic signals. The mitogen-activated protein kinase (MAPK) pathway plays a vital role in controlling cell proliferation and differentiation in the developing CNS, since the MARK pathway regulates various physiological signaling inputs and serves as a relay route from the cell surface to the nucleus . The breast cancer suppressor protein (BRCA1) associated protein (BRAP) is a cytoplasmic protein which regulates nuclear targeting by retaining proteins with a nuclear localization signal in the cytoplasm . BRAP, previously known as impedes mitogenic propagation (IMP), is a Ras effector and a negative regulator of the MAPK scaffold protein kinase suppressor of Ras 1 (KSR1) . Thus, BRAP is regarded as a threshold modulator that controls the sensitivity of MAPK signaling, allowing adaptations, and maintains homeostasis in a complex tissue environment . Such threshold regulation is specifically important in CNS development, during which each progenitor cell has a unique environment and its fate choices are precisely correlated to its spatial location. In addition, BRAP is proposed to be a modulator that associates with Skp1-Cullin1-F-box protein (SCF) complex and controls TNF-a-induced nuclear factor-kappaB (NF-κB) nuclear translocation; while NF-κB is critical for the expression of multiple genes involved in inflammatory responses and cellular survival . BRAP is also an anchor protein for p21 through direct interaction . These lines of evidence suggest that BRAP can control different kinds of intracellular signals. Interestingly, recent genetic analyses have revealed that BRAP is associated with several human disorders caused by inflammatory dysfunction, including myocardial infarction, carotid atherosclerosis and central obesity -.
Given the close link between developmental neuroinflammation and SZ, we hypothesized that the BRAP gene may influence the genesis of SZ. In this work, we aimed to evaluate the association of BRAP polymorphisms with SZ in a two-stage study. In the discovery stage, we derived relevant data from our GWAS data , involving six SNPs in 768 SZ cases and 1348 healthy controls. In the validation stage, we replicated the most significant SNPs in 1957 cases and 1509 controls. Finally, we sought to explore the potential role the associated SNP(s) using cis-eQTL (expression quantitative trait loci) analysis via online datasets.
Materials and Methods
All participants were unrelated Han Chinese recruited from the North of China. The GWAS sample consisted of 768 unrelated subjects with SZ (360 males and 408 females) and 1348 control subjects (658 males and 690 females). For validation, an independent sample consisting of 1957 cases (1037 males and 920 females) and 1509 controls (360 males and 1149 females) was recruited from northern China. The consensus diagnoses were made by at least two experienced psychiatrists according to the Diagnosis and Statistical Manual of Mental Disorders Fourth Edition (DSM-IV) criteria for schizophrenia. None of the patients had severe medical complications. Healthy controls were recruited from communities with simple non-structured interview performed by psychiatrists, who excluded individuals with history of mental health and neurological diseases.
The study was approved by the Medical Research Ethics Committee of the Institute of Mental Health, Peking University. All participants enrolled in the study signed written informed consent.
Peripheral blood samples were collected from all subjects. Genomic DNA was extracted using the Qiagen QIAamp DNA Mini Kit. In the screening stage, we derived genotypes of six SNPs in BRAP from our GWAS data , involving rs1544396, rs11065987, rs10744774, rs17628828, rs3782886, and rs601663. In the validation stage, the genotypes of rs1544396, rs3782886, and rs601663 were determined using the Sequenom MassARRAY system (Sequenom iPLEX). 1957 cases and 1059 controls were successfully genotyped, the calling rate being 99.91%. Data from this study are freely available upon request.
Genetic association tests were analyzed using PLINK 1.07 , and 10000 permutations were used to achieve the adjusted P values. Linkage disequilibrium (LD) among the markers was plotted with Haploview . The pie chart of the worldwide distribution of rs3782886 was created from HGDP Selection Browser . Genetic power was estimated by PS . We performed the cis-eQTL using Genevar 3.2  in two Asian populations from HapMap3 project: CHB (80 Han Chinese from Beijing, China) and JPT (82 Japanese in Tokyo, Japan) , which analyzed the association of rs3782886 genotypes with neighboring genes within 1 Mb distance. Protein-protein interaction network was built via STRING v9.1 . Finally, we predicted the function of rs3782886 via the on-line tool F-SNP (http://compbio.cs.queensu.ca/F-SNP/) .
Genotypic distributions of the six SNPs did not deviate from Hardy-Weinberg equilibrium (HWE) in either the patient group or the control group (P > 0.05). The BRAP maps on chromosome 12q24, and the six SNPs span across the gene (Figure 1). The LD plot among the SNPs is shown in Figure 2. Three SNPs were found to be associated with SZ, including rs1544396 (P = 2.75E-5, adjusted P = 2.00E-4, OR = 1.63), rs3782886 (P = 2.31E-6, adjusted P = 1.00E-4, OR = 0.67), and rs601663 (P = 0.0047, adjusted P = 0.025, OR = 1.56) (Table 1). The most significant SNP, rs3782886, was replicated in the validation data (P = 1.43E-6, adjusted P = 1.00E-4, OR = 0.73), while the other two SNPs failed to be replicated. Thus, rs3782886 was considered be consistently associated with SZ across the two samples. The allele frequencies of rs3782886 across the world is shown in Figure 3.
cis-eQTL analysis and SNP function prediction
Several genes are located near the BRAP gene on chromosome 12q24, including aldehyde dehydrogenase 2 family (ALDH2) and myosin regulatory light chain 2 (MYL2) (Figure 4). Of note is that BRAP is located at the upper stream of both ALDH2 and MYL2. The cis-eQTL analysis detected two associated genes (Figure 5). There is an inverse correlation between the dosage of rs3782886 A-allele and ALDH2 mRNA levels in the Han Chinese population (Ppermutation = 0.0039), with the Spearman's rank correlation coefficient (ρ) being –0.315; while, the dosage of rs3782886 A-allele correlates positively with the MYL2 mRNA levels in the Japanese population (Ppermutation < 1.0E-4, ρ = 0.421, Figure 6). SNP rs3782886 was predicted to influence several transcriptional factors, and its functional significance (FS) score is 0.5, suggesting a possible role of in gene function (the median FS score for neutral SNPs is 0.176, whereas, for disease-related SNPs, the median rises to 0.5, Figure S1) . rs3782886 is predicted to interact directly with several other proteins (Figure 7).
Two genes, ALDH2 and MYL2, were associated with rs3782886 genotypes in the Chinese (CHB) and Japanese (JPT) populations, respectively.
ρ: Spearman's rank correlation coefficient. Pemp: Adjusted P-value from 10000 permutation tests.
SZ is one of most disabling psychoses, affecting virtually all brain functions. Although the cause of this disorder remains elusive, by far the vast majority of evidence points to the neurodevelopmental model in which abnormalities of early brain development predispose to future onset of SZ. And it has been postulated that immune system exerts an effect on normal CNS development while abnormal immune reaction may adversely impede this process. Empirical evidence indicates that immune-mediated disruption of early brain development contributes to the precipitation of later psychotic disease . Salient examples include roles of viral infection  and oxidative stress  in the genesis of SZ. Thus, both neuroprogression and neuroinflammation process could exert pivotal roles in human CNS maturation and functioning.
SZ is a complex genetic disease with an estimated heritability of approximately 80–85% . Genetic epidemiology and population genetics suggest that a spectrum of genetic risk variants from a host of candidate genes contribute to the disorder. Mounting work has revealed lots of insight into the genetic underpinnings of SZ, nevertheless, the larger bulk of disease heritability remains unaccounted for . Part of the problem may lie in the fact that most efforts have been focusing on the dopamine system in the search of its genetic basis, with little emphasis on neurodevelopmental pathway genes.
BRAP is a polytrophic molecule which plays vital roles in several signaling pathways. First, it is a threshold regulator in the MAPK pathway, while the MAPK cascade forms the backbone of cell signaling and has an extraordinarily broad impact in cell growth and differentiation. Second, BRAP also acts as a key mediator of inflammatory cascades by regulating NF-κB nuclear translocation , while the immune system participates in the neural progression. Thus, BRAP acts more as a molecular buffer to maintain homeostasis of a cell’s response to the environment , and is critical to cerebral development as well as brain function. Apart from KSR1 and BRCA1, protein-protein network analysis suggests several other interacting partners. Of note is ubiquitin C (UBC), a critical ‘switchboard’ node which was implicated in SZ by a mRNA expression study .
BRAP is located on chromosome 12q24, one of the SZ-related chromosomal fragile sites  harboring several candidate genes for SZ, including nitric oxide synthase 1 (NOS1)  and D-amino acid oxidase (DAO) . SNP rs11066001 lies in intron 3, while rs3782886 is a synonymous SNP in exon 5. Of note is that rs11066001 is in LD with rs3782886, with R2 of 0.891 and D` of 1 in our data. Some studies reported that two SNPs—rs11066001 and rs3782886—in BRAP were significantly associated with coronary artery disease in Asian populations , , and metabolic syndrome in the Han Chinese population . Another study reported the association of rs11066001 with carotid atherosclerosis in Taiwanese . Notably, all these studies were conducted in Asian populations. According to HapMap data, these two SNPs are nonpolymorphic in European (CEU) and African (YRI) population, therefore rs3782886 seems likely to be present chiefly in Asian populations (Figure 3). However, another SNP (rs11065987) in BRAP was reported to be associated with metabolic syndrome in European Americans and African Americans , suggesting genetic heterogeneity of BRAP across diverse populations.
In this work, we genotyped and analyzed six polymorphisms within BRAP, detecting three significant SNPs: rs1544396, rs3782886, and rs601663. The most significant SNP, rs3782886, was replicated in the validation data, while the other two SNPs failed to be replicated. Thus, our data strongly support the association of rs3782886 with SZ in Han Chinese population, with the mutant G-allele conferring a protective effect against SZ disease (OR = 0.68).
The cis-eQTL analysis revealed that genotypes of rs3782886 were associated with mRNA levels of two nearby genes—ALDH2 and MYL2—in the Chinese and Japanese population, respectively. rs3782886 was predicted to influence several transcriptional factors . It is noteworthy that rs3782886 is located at the upper stream of both ALDH2 and MYL2. Considering that the upper stream of a gene harbors vital regulating elements for transcription, it is plausible that such a cis-regulating effect may exist.
The ALDH2 protein is the mitochondrial isoform of aldehyde dehydrogenase, which metabolizes acetaldehyde and governs the detoxification of acetaldehyde formed following alcohol consumption and the ultimate elimination of alcohol from the body . In addition to metabolizing acetaldehyde, aldehyde dehydrogenase is also involved in dopamine and serotonin metabolism . Genetic studies suggested that ALDH2 was associated with alcohol dependence , bipolar I disorder , and cardiovascular diseases . Interestingly, both BRAP and ALDH2 were found to be associated with coronary artery disease in the Japanese .
MYL2 encodes the myosin regulatory light chain that specifically binds to myosin heavy chains, constituting one of the most important components of sarcomere filaments . MYL2 also mediates the AMP-activated protein kinase (AMPK) pathway which regulates energy homeostasis in eukaryotes . This gene was found to be linked with multiple cardiovascular diseases  and drinking behavior .
Considering the involvement of dopamine and serotonin in SZ and the comorbidity of SZ with alcohol dependence and cardiovascular diseases, it is plausible that these two genes (ALDH2 and MYL2) may be related to SZ. The relationship of rs3782886 genotypes with BRAP mRNA is unknown because of lacking the BRAP data in the dataset , nevertheless, a link between them may possibly exist since rs3782886 can affect the transcriptional activity of the gene . Together, these data suggest that rs3782886 may represent effects beyond the BRAP gene, possible through direct cis-regulation or through proxy effect by in LD with the truly regulatory SNPs.
Assuming a disease allele frequency of 0.20, and a type I error of 0.05, our sample has a 93.8% power to detect a variant with an odds ratios of 0.73 . The major limitation may be the selection of SNPs were not unbiased, but relied on the design of the commercial array. Due to constraint of fund, we were not able to biologically validate the role of rs3782886 on the gene function, which is warranted to be explored in the future.
In conclusion, our findings suggest BRAP as a candidate gene for SZ the Han Chinese population. Our data provide novel genetic evidence for the involvement of developmental and/or neuroinflammatory cascades in SZ.
We would sincerely thank the patients, their families and the healthy volunteers for their participation, and all the medical staff involved in specimen collecting.
Conceived and designed the experiments: DZ WY. Performed the experiments: YX TL LT GW YC. Analyzed the data: FZ CL. Contributed reagents/materials/analysis tools: CL GQ GY GZ ZZ XZ HY CJ YR LW JY JW ZW WZ ZC. Wrote the paper: FZ WY.
- 1. Messias EL, Chen CY, Eaton WW (2007) Epidemiology of schizophrenia: review of findings and myths. Psychiatr Clin North Am 30: 323–338.
- 2. Murray RM, Lewis SW (1987) Is schizophrenia a neurodevelopmental disorder? Br Med J (Clin Res Ed) 295: 681–682.
- 3. Weinberger DR (1987) Implications of normal brain development for the pathogenesis of schizophrenia. Arch Gen Psychiatry 44: 660–669.
- 4. van Os J, Kapur S (2009) Schizophrenia. Lancet 374: 635–645.
- 5. Owen MJ, O'Donovan MC, Thapar A, Craddock N (2011) Neurodevelopmental hypothesis of schizophrenia. Br J Psychiatry 198: 173–175.
- 6. Meyer U (2013) Developmental neuroinflammation and schizophrenia. Prog Neuropsychopharmacol B_ol Psychiatry 42: 20–34.
- 7. Mehler MF, Kessler JA (1997) Hematolymphopoietic and inflammatory cytokines in neural development. Trends Neurosci 20: 357–365.
- 8. Wellbrock C, Karasarides M, Marais R (2004) The RAF proteins take centre stage. Nat Rev Mol Cell Biol 5: 875–885.
- 9. Li S, Ku CY, Farmer AA, Cong YS, Chen CF, et al. (1998) Identification of a novel cytoplasmic protein that specifically binds to nuclear localization signal motifs. J Biol Chem 273: 6183–6189.
- 10. Matheny SA, Chen C, Kortum RL, Razidlo GL, Lewis RE, et al. (2004) Ras regulates assembly of mitogenic signalling complexes through the effector protein IMP. Nature 427: 256–260.
- 11. Lanctot AA, Peng CY, Pawlisz AS, Joksimovic M, Feng Y (2013) Spatially dependent dynamic MAPK modulation by the Nde1-Lis1-Brap complex patterns mammalian CNS. Dev Cell 25: 241–255.
- 12. Takashima O, Tsuruta F, Kigoshi Y, Nakamura S, Kim J, et al. (2013) Brap2 regulates temporal control of NF-kappaB localization mediated by inflammatory response. PLoS One 8: e58911.
- 13. Asada M, Ohmi K, Delia D, Enosawa S, Suzuki S, et al. (2004) Brap2 functions as a cytoplasmic retention protein for p21 during monocyte differentiation. Mol Cell Biol 24: 8236–8243.
- 14. Ozaki K, Sato H, Inoue K, Tsunoda T, Sakata Y, et al. (2009) SNPs in BRAP associated with risk of myocardial infarction in Asian populations. Nat Genet 41: 329–333.
- 15. Liao YC, Wang YS, Guo YC, Ozaki K, Tanaka T, et al. (2011) BRAP Activates Inflammatory Cascades and Increases the Risk for Carotid Atherosclerosis. Mol Med 17: 1065–1074.
- 16. Avery CL, He Q, North KE, Ambite JL, Boerwinkle E, et al. (2011) A phenomics-based strategy identifies loci on APOC1, BRAP, and PLCG1 associated with metabolic syndrome phenotype domains. PLoS Genet 7: e1002322.
- 17. Yue WH, Wang HF, Sun LD, Tang FL, Liu ZH, et al. (2011) Genome-wide association study identifies a susceptibility locus for schizophrenia in Han Chinese at 11p11.2. Nat Genet 43: 1228–1231.
- 18. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 81: 559–575.
- 19. Barrett JC, Fry B, Maller J, Daly MJ (2005) Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 21: 263–265.
- 20. Pickrell JK, Coop G, Novembre J, Kudaravalli S, Li JZ, et al. (2009) Signals of recent positive selection in a worldwide sample of human populations. Genome Res 19: 826–837.
- 21. Dupont WD, Plummer WD (1998) Power and sample size calculations for studies involving linear regression. Control Clin Trials 19: 589–601.
- 22. Yang TP, Beazley C, Montgomery SB, Dimas AS, Gutierrez-Arcelus M, et al. (2010) Genevar: a database and Java application for the analysis and visualization of SNP-gene associations in eQTL studies. Bioinformatics 26: 2474–2476.
- 23. Stranger BE, Montgomery SB, Dimas AS, Parts L, Stegle O, et al. (2012) Patterns of cis regulatory variation in diverse human populations. PLoS Genet 8: e1002639.
- 24. Franceschini A, Szklarczyk D, Frankild S, Kuhn M, Simonovic M, et al. (2013) STRING v9.1: protein-protein interaction networks, with increased coverage and integration. Nucleic Acids Res 41: D808–815.
- 25. Lee PH, Shatkay H (2008) F-SNP: computationally predicted functional SNPs for disease association studies. Nucleic Acids Res 36: D820–824.
- 26. Lee PH, Shatkay H (2009) An integrative scoring system for ranking SNPs by their potential deleterious effects. Bioinformatics 25: 1048–1055.
- 27. Kneeland RE, Fatemi SH (2013) Viral infection, inflammation and schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry 42: 35–48.
- 28. Wood SJ, Yucel M, Pantelis C, Berk M (2009) Neurobiology of schizophrenia spectrum disorders: the role of oxidative stress. Ann Acad Med Singapore 38: 396–396.
- 29. Cardno AG, Gottesman II (2000) Twin studies of schizophrenia: from bow-and-arrow concordances to star wars Mx and functional genomics. Am J Med Genet 97: 12–17.
- 30. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, et al. (2009) Finding the missing heritability of complex diseases. Nature 461: 747–753.
- 31. Lee SA, Tsao TT, Yang KC, Lin H, Kuo YL, et al. (2011) Construction and analysis of the protein-protein interaction networks for schizophrenia, bipolar disorder, and major depression. BMC Bioinformatics 12 Suppl 13S20.
- 32. Demirhan O, Tastemir D, Sertdemir Y (2006) Chromosomal fragile sites in schizophrenic patients. Genetika 42: 985–992.
- 33. Reif A, Herterich S, Strobel A, Ehlis AC, Saur D, et al. (2006) A neuronal nitric oxide synthase (NOS-I) haplotype associated with schizophrenia modifies prefrontal cortex function. Mol Psychiatry 11: 286–300.
- 34. Verrall L, Burnet PW, Betts JF, Harrison PJ (2010) The neurobiology of D-amino acid oxidase and its involvement in schizophrenia. Mol Psychiatry 15: 122–137.
- 35. Hinohara K, Ohtani H, Nakajima T, Sasaoka T, Sawabe M, et al. (2009) Validation of eight genetic risk factors in East Asian populations replicated the association of BRAP with coronary artery disease. J Hum Genet 54: 642–646.
- 36. Wu L, Xi B, Hou D, Zhao X, Liu J, et al. (2013) The single nucleotide polymorphisms in BRAP decrease the risk of metabolic syndrome in a Chinese young adult population. Diabetes & vascular disease research : official journal of the International Society of Diabetes and Vascular Disease 10: 202–207.
- 37. Zhang Y, Ren J (2011) ALDH2 in alcoholic heart diseases: molecular mechanism and clinical implications. Pharmacol Ther 132: 86–95.
- 38. Chen CC, Lu RB, Chen YC, Wang MF, Chang YC, et al. (1999) Interaction between the functional polymorphisms of the alcohol-metabolism genes in protection against alcoholism. Am J Hum Genet 65: 795–807.
- 39. Lee SY, Chen SL, Chang YH, Chu CH, Huang SY, et al. (2012) The ALDH2 and 5-HT2A genes interacted in bipolar-I but not bipolar-II disorder. Prog Neuropsychopharmacol B_ol Psychiatry 38: 247–251.
- 40. Takeuchi F, Yokota M, Yamamoto K, Nakashima E, Katsuya T, et al. (2012) Genome-wide association study of coronary artery disease in the Japanese. Eur J Hum Genet 20: 333–340.
- 41. Wadgaonkar R, Shafiq S, Rajmanickam C, Siddiqui MA (1993) Interaction of a conserved peptide domain in recombinant human ventricular myosin light chain-2 with myosin heavy chain. Cell Mol Biol Res 39: 13–26.
- 42. Thaiparambil JT, Eggers CM, Marcus AI (2012) AMPK regulates mitotic spindle orientation through phosphorylation of myosin regulatory light chain. Mol Cell Biol 32: 3203–3217.
- 43. Weterman MA, Barth PG, van Spaendonck-Zwarts KY, Aronica E, Poll-The BT, et al. (2013) Recessive MYL2 mutations cause infantile type I muscle fibre disease and cardiomyopathy. Brain 136: 282–293.
- 44. Yang X, Lu X, Wang L, Chen S, Li J, et al. (2013) Common variants at 12q24 are associated with drinking behavior in Han Chinese. Am J Clin Nutr 97: 545–551.