Whole-genome association studies of alcoholism with loci linked to schizophrenia susceptibility

  • Junghyun Namkung1,

    Affiliated with

    • Youngchul Kim2 and

      Affiliated with

      • Taesung Park1, 2Email author

        Affiliated with

        BMC Genetics20056(Suppl 1):S9

        DOI: 10.1186/1471-2156-6-S1-S9

        Published: 30 December 2005

        Abstract

        Background

        Alcoholism is a complex disease. There have been many reports on significant comorbidity between alcoholism and schizophrenia. For the genetic study of complex diseases, association analysis has been recommended because of its higher power than that of the linkage analysis for detecting genes with modest effects on disease.

        Results

        To identify alcoholism susceptibility loci, we performed genome-wide single-nucleotide polymorphisms (SNP) association tests, which yielded 489 significant SNPs at the 1% significance level. The association tests showed that tsc0593964 (P-value 0.000013) on chromosome 7 was most significantly associated with alcoholism. From 489 SNPs, 74 genes were identified. Among these genes, GABRA1 is a member of the same gene family with GABRA2 that was recently reported as alcoholism susceptibility gene.

        Conclusion

        By comparing 74 genes to the published results of various linkage studies of schizophrenia, we identified 13 alcoholism associated genes that were located in the regions reported to be linked to schizophrenia. These 13 identified genes can be important candidate genes to study the genetic mechanism of co-occurrence of both diseases.

        Background

        Alcoholism is a complex disease that tends to run in families. It has been reported that alcoholism is accompanied by many other psychiatric disorders, including schizophrenia [1]. The co-occurrence of these psychiatric disorders and alcoholism can be explained in several ways. First, patients may use alcohol to relieve the symptom of mental diseases. Second, disruption of neurochemical systems may cause psychiatric disorders and alcoholism simultaneously. Several molecules including serotonin, neuropeptide Y, and dopamine have been studied to explore this hypothesis. Third, the genes responsible for these diseases may transmit together because they are closely linked on a chromosome and co-segregate without any functional relationship.

        Linkage or association study of polymorphisms may provide useful information on the genetic mechanism of co-occurrence of psychiatric disorders and alcoholism. Among psychiatric disorders, schizophrenia is also known to be affected by multiple genes [2]. Schizophrenia patients with substance abuse problems, especially alcoholics, are clinically important because they usually have a poor prognosis [3, 4].

        In this study, we are interested in determining whether there is any common genetic factor that may increase the susceptibility of alcoholism and schizophrenia simultaneously. Although linkage analysis has been a successful choice for genetic analysis of Mendelian diseases, association analysis has shown to have greater power than linkage analysis when used to detect genes with modest effect on disease [5]. Thus, we conduct genome-wide association tests to find alcoholism susceptibility loci by analyzing the Collaborative Study on the Genetics of Alcoholism (COGA) data from the Genetic Analysis Workshop 14 (GAW 14). Because COGA data were collected from multiplex families, we consider only the approaches that take advantage of the whole pedigree data.

        In our analysis, we primarily use the pedigree disequilibrium test (PDT) [6]. The PDT is an extension of the transmission disequilibrium test (TDT) [7], and it has been widely applied to the analysis of large pedigree data including multiple nuclear family or sibling pairs. We test association for both alleles and genotypes using the PDT. Additionally, we apply the generalized estimating equations (GEE) approach [8] to test the association for each genotype with the disease while adjusting for some phenotypic covariates. By using both the PDT and GEE approaches we obtain candidate markers for alcoholism susceptibility and then compare them with the loci that are reported to be linked to schizophrenia.

        Methods

        Disease affection status was defined by using both ALDX1 (DSM-III-R+Feighner) and ALDX2 (DSM-IV). A sample with ALDX1 = 5 or ALDX2 = 5 was defined as the affected, and ALDX1 = 1 or ALDX2 = 1 as the unaffected. Others treated as non-informative data. With these criteria, 668 were classified as "affected" and 285 as "unaffected" from 1,614 individuals.

        Linkage analysis

        For the comparison of association with the linkage analysis, we first performed nonparametric linkage (NPL) analysis for microsatellites. Due to computational problems, NPL analysis for single-nucleotide polymorphisms (SNPs) was not performed. The linkage analysis for microsatellites was performed by using NPL procedure of SIMWALK2 software [9]. To transform GAW14 data into input format for SIMWALK2 software, we used MEGA2 software [10].

        PDT

        To test for association of microsatellites and SNPs with disease, we used the PDT. Because the PDT is a family-based association test, it can avoid problems of false positives caused by population stratification. There are two types of the PDT available: allele-based PDT and genotype-based PDT (genotype-PDT). For the allele-based PDT, two test statistics are commonly used: sum-PDT and average-PDT. The sum-PDT gives more weight to families with a larger number of multiple affected individuals. In contrast, the average-PDT gives equal weight to all families [11]. The genotype-PDT tests for association between genotypes and disease. This is more powerful than the allele-based PDT when the genetic effect of an allele is dominant or recessive rather than additive. The genotype-PDT can test multilocus effects without the ambiguities associated with haplotype analysis, and allow for testing interactions among markers [12]. We used the genotype-PDT to test for association between disease and genotypes at a single locus as well as at multiple loci. All the tests were conducted by using PDT 5.1 software.

        GEE

        The GEE approach accounts for familial correlation in the analysis and allows for adjustment for covariates such as sex and age in selecting significant genotypes of SNPs. For the binary response of disease affection status, the model is given by

        Logit [pr(D ij )] = β 0 + β ageage ij + β sexsex ij + β genogenotype ij , where

        D ij is the affection status of j th individual in i th family

        (affected = 1, unaffected = 0)

        We conducted this analysis using R software http://​www.​r-project.​org/​.

        Gene finding

        After identifying SNPs associated with alcoholism, we obtained gene information from the dbSNP website http://​www.​ncbi.​nlm.​nih.​gov/​SNP/​. The additional information of genes such as chromosomal location and functional annotation was obtained from the SOURCE website http://​source.​stanford.​edu/​. Then, we compared the genes and regions in which they are located to published results of various linkage studies of schizophrenia.

        Results

        Preliminary analysis using microsatellites

        Before performing the association analysis using large number of SNPs, we conducted linkage and association analyses for microsatellite markers. At first, a chromosome-wide association test was performed using two PDT statistics: average-PDT and sum-PDT. At the 5% significance level (α = 0.05) for either the sum-PDT or average-PDT, this preliminary test shows that 6 microsatellites from chromosomes 6 and 8 tend to have associations with alcoholism. Among them, two microsatellites are located adjacent to the schizophrenia susceptibility region [2, 13]: 1) D6S474 (6q21) with p-values 0.0182 for the sum-PDT and 0.0144 for the average-PDT, 2) D8S1106 (8p21 SCZD6) with p-value 0.0155 for the sum-PDT.

        To test for linkage of markers to alcoholism, we performed NPL analysis. Chromosome 7 showed two peaks of moderate linkage disequilibrium with alcoholism at D7S673 (30.1 cM) with -log p-value 2.077 (NPL-PAIR) and D7S820 (107.5 cM) with 2.041.

        Association test for SNPs

        From 15,878 autosomal SNPs, we first selected SNPs that have p-values smaller than 0.05 for both allele based PDTs. One hundred and ninety SNPs had p-values less than 0.01 and 18 SNPs had p-values less than 0.001 by either of the two PDTs. Using the genotype-PDT we identified 138 and 16 SNPs showing significant association at α = 0.01 and 0.001, respectively. Among 138 SNPs, 94 also showed significant associations for the allele-based PDT at α = 0.01.

        SNP tsc0593964 on chromosome 7 showed the most significant association with alcoholism (p-value = 0.000013, genotype-PDT). It is located about 2.3 cM away from the modest NPL peak of D7S673 found in the linkage analysis of microsatellites (Figure 1).
        http://static-content.springer.com/image/art%3A10.1186%2F1471-2156-6-S1-S9/MediaObjects/12863_2005_Article_270_Fig1_HTML.jpg
        Figure 1

        Linkage for microsatellites and association for SNPs on chromosome 7.

        As an alternative, we also applied the GEE approach, which considers familial correlation. From this analysis, only 65 SNPs showed p-values less than 0.01. rs1262129 appeared to be most significant (p-value = 0.000068). Among the 65 significant SNPs, two SNPs were also significant for the genotype-PDT: tsc0253130 (p-value = 0.0007) and tsc0834636 (p-value = 0.0054). The number of significant SNPs is much smaller than that of PDT. It is probably due to the lack of power caused by ignoring genetic inheritance information in the GEE approach.

        In addition, we performed multipoint association analysis using the multipoint genotype-PDT. The multipoint genotype-PDT tests for association of multiple loci without haplotyping. We conducted 2-point, 3-point, and 5-point analysis for the adjacent markers. We obtained 98 SNP pairs from the 2-point analysis, and 42 SNP triple sets from the 3-point analysis at the 1% significance level. The 5-point analysis resulted in no significant marker sets. Some identified SNP sets did not contain any SNP that showed significant association from the single-locus test. This underscores the importance of interactions among multiple markers to search for candidate disease genes.

        Using the dbSNP and the SOURCE database website, we selected intragenic SNPs among those associated with alcoholism. A total of 74 known genes were found to contain the significant SNPs after hypothetical genes or open reading frames were excluded. Two of these genes, GABRA1 (tsc0325674, p-value = 0.0035 by the average-PDT) and CHRNA3 (rs1878399, p-value = 0.0083 by the average-PDT) are neurotransmitter receptors; NTRK2 (tsc0656804, p-value = 0.0041 by the average-PDT) has function of neurogenesis. Of note, GABRA1 is a member of the same gene family as GABRA2, which was reported recently to be associated with alcoholism [14].

        Finally, we compared the chromosomal location of the 74 genes with the schizophrenia susceptibility regions [2, 13]. The regions were previously identified by several linkage studies of schizophrenia. The comparison revealed that 13 genes were located in schizophrenia susceptibility region (Table 1). These genes and SNPs on them can be putative markers responsible for both alcoholism and schizophrenia susceptibility.
        Table 1

        The SNPs showing significant association located in major schizophrenia susceptibility candidate regions

        Chr

        SNP ID

        Gene symbol

        Cytoband

        PDT p-value

        Reported schizophrenia regione

        1

        rs908857

        DUSP10

        1q41

        0.0043a

        1q32.2-q41

         

        rs1053074-rs10594

        KCNJ10

        1q22-q23

        0.0055c

        1q21-22

         

        tsc1457991-tsc1254625

        PBX1

        1q23

        0.001c

        1q22-23

        5

        rs1229708

        SPOCK

        5q31

        0.009a

        5q31

         

        tsc0935735

        FER

        5q21

        0.0088b

        5q21-31

        6

        rs1498426

        CDKAL1

        6p22.3

        0.006b

        6p22-24

         

        rs1022092

        SLC16A10

        6q21-q22

        0.0039b

         
         

        tsc0046065

        PREP

        6q22

        0.0022a

        6q21-q22.3

         

        tsc0615608

        AIM1

        6q21

        0.0087c

         
         

        tsc0253130

        TCBA1

        6q21

        0.0003b

        (0.009)d

         

        8

        tsc0149489-tsc0514918-tsc0529734

        PTK2B

        8p21.1

        0.0056c

        8p21-22

         

        rs898249-rs900267-rs310319

        BIN3

        8p21.3

        0.0053c

         

        10

        rs729245

        CACNB2

        10p12

        0.0009a

        10p11-15

        aAverage-PDT p-value

        bSum-PDT p-value

        cGenotype-PDT p-value

        dGEE p-value

        eData were brought from Michael et al.'s review article [2]. SNPs listed are located in genetic regions and showed allelic or genotypic association at 1% significance level. The schizophrenia susceptibility regions were identified by several independent linkage studies.

        Summary and Discussion

        Through the analysis using microsatellites, we obtained rough regions showing modest linkage or association with alcoholism. Later, we found that the most significant SNP, tsc0593964, was located near the modest linkage peak on chromosome 7.

        We conducted the genome-wide association analysis for large number of SNPs by using different association methods such as the allele-based PDT, genotype-based PDT for single locus or for multiple loci, and GEE. Each method resulted in a different number of significant markers. Thus, the results should be interpreted carefully. We think it would be important to compare the results in a more systematic way by using simulation studies in the future.

        With additional biological information such as gene names and functions, the list of selected candidate markers may be quite useful for the study of complex disease, even though the 1% significance level is not stringent enough for the genome-wide tests. We searched for gene information of the selected SNPs and found that some of the genes are related to the neurochemical system.

        The evidence for significant co-morbidity of substance abuse, especially alcoholism, and schizophrenia is very robust [4]. The genetic relationship between two diseases has not yet been explained. We think our findings of the 13 genes, which are associated with alcoholism and also located in schizophrenia linkage regions, can be helpful to study genetic factors responsible for both diseases.

        Abbreviations

        COGA: 

        Collaborative Study on the Genetics of Alcoholism

        GAW14: 

        Genetic Analysis Workshop 14

        GEE: 

        Generalized estimating equation

        NPL: 

        Nonparametric linkage

        PDT: 

        Pedigree disequilibrium test

        SNP: 

        Single-nucleotide polymorphism

        TDT: 

        Transmission disequilibrium test

        Declarations

        Acknowledgements

        The authors thank for two anonymous reviewers for their helpful comments. This work was supported by the National Research Program of Korea Science and Engineering Foundation.

        Authors’ Affiliations

        (1)
        Bioinformatics Program, Seoul National University
        (2)
        Department of Statistics, Seoul National University

        References

        1. Ross HE, Glaser FB, Germanson T: The prevalence of psychiatric disorders in patients with alcohol and other drug problems. Arch Gen Psychiatry 1988, 45:1023–1031.PubMed
        2. Michael CO, Nigel MW, Michael JO: Recent advances in the genetics of schizophrenia. Hum Mol Genet 2003,12(review issue 2):R125–R133.
        3. Crome IB: Alcoholism and schizophrenia: a challenge for treatment and research. Addiction 2000, 95:1859–1860.View ArticlePubMed
        4. Soyka M: Alcoholism and schizophrenia. Addiction 2000, 95:1613–1618.View ArticlePubMed
        5. Risch N, Merikangas K: The future of genetic studies of complex human diseases. Science 1996, 273:1516–1517.View ArticlePubMed
        6. Martin ER, Monks SA, Warren LL, Kaplan NL: A test for linkage and association in general pedigrees: the pedigree disequilibrium test. Am J Hum Genet 2000, 67:146–154.View ArticlePubMed
        7. Spielman RS, Ralph EM, Warren JE: Transmission test for linkage disequilibrium: the insulin gene region and insulin–dependent diabetes mellitus (IDDM). Am J Hum Genet 1993, 52:506–516.PubMed
        8. Zeger SL, Liang KY: Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986, 42:121–130.View ArticlePubMed
        9. Sobel E, Lange K: Descent graphs in pedigree analysis: applications to haplotyping, location scores, and marker sharing statistics. Am J Hum Genet 1996, 58:1323–1337.PubMed
        10. Mukhopadhyay N, Almasy L, Schroeder M, Mulvihill WP, Weeks DE: Mega2, a data–handling program for facilitating genetic linkage and association analyses. Am J Hum Genet 1999, 65:A436.
        11. Martin ER, Bass MP, Kaplan NL: Correction for a potential bias in the pedigree disequilibrium test. Am J Hum Genet 2001, 68:1065–1067.View ArticlePubMed
        12. Martin ER, Bass MP, Gilbert JR, Pericak–Vance MA, Hauser ER: Genotype–based association test for general pedigrees: the genotype–PDT. Genet Epidemiol 2003, 25:203–213.View ArticlePubMed
        13. Blouin JL, Dombroski BA, Nath SK, Lasseter VK, Wolyniec PS, Nestadt G, Thornquist M, Ullrich G, McGrath J, Kasch L, Lamacz M, Thomas MG, Gehrig C, Radhakrishna U, Snyder SE, Balk KG, Neufeld K, Swartz KL, DeMarchi N, Papadimitriou GN, Dikeos DG, Stefanis CN, Chakravarti A, Childs B, Housman DE, Kazazian HH, Antonarakis SE, Pulver AE: Schizophrenia susceptibility loci on chromosomes 13q32 and 8p21. Nat Genet 1998, 20:70–73.View ArticlePubMed
        14. Edenberg HJ, Dick DM, Xuei X, Tian H, Almasy L, Bauer LO, Crowe RR, Goate A, Hesselbrock V, Jones K, Kwon J, Li TK, Nurnberger JI Jr, O'Connor SJ, Reich T, Rice J, Schuckit MA, Porjesz B, Foroud T, Begleiter H: Variations in GABRA2, encoding the alpha 2 subunit of the GABA(A) receptor, are associated with alcohol dependence and with brain oscillations. Am J Hum Genet 2004, 74:705–714.View ArticlePubMed

        Copyright

        © Namkung et al 2005

        This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://​creativecommons.​org/​licenses/​by/​2.​0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

        Advertisement