Association study of NRG1 DTNBP1 RGS4 G72G30 and PIP5K2A with schizophrenia and symptom severity in a Hungarian sample.код для вставкиСкачать
RESEARCH ARTICLE Neuropsychiatric Genetics Association Study of NRG1, DTNBP1, RGS4, G72/G30, and PIP5K2A With Schizophrenia and Symptom Severity in a Hungarian Sample Janos M. Rethelyi,1* Steven C. Bakker,2 Patrıcia Polgar,1 Pal Czobor,1,3 Eric Strengman,4 Peter I. Pasztor,5 Rene S. Kahn,2 and Istvan Bitter1 1 Department of Psychiatry and Psychotherapy, Semmelweis University, Budapest, Hungary 2 Department of Psychiatry, The Rudolf Magnus Institute of Neuroscience, University Medical Center Utrecht, Utrecht, the Netherlands 3 Nathan Kline Institute for Psychiatric Research, Orangeburg, New York Complex Genetics Section, Department of Biomedical Genetics, University Medical Center Utrecht, Utrecht, the Netherlands 4 5 Department of Psychiatry, St. John’s Hospital, Budapest, Hungary Received 19 December 2008; Accepted 29 September 2009 Genetic association studies have yielded extensive but frequently inconclusive data about genetic risk factors for schizophrenia. Clinical and genetic heterogeneity are possible factors explaining the inconsistent findings. The objective of this study was to test the association of commonly incriminated candidate genes with two clinically divergent subgroups, non-deficit (SZ-ND) and deficit-schizophrenia (SZ-D), and symptom severity, in order to test for replication of previously reported results. A homogeneous sample of 280 schizophrenia patients and 230 healthy controls of Hungarian, Caucasian descent were genotyped for polymorphisms in schizophrenia candidate genes NRG1, DTNBP1, RGS4, G72/G30, and PIP5K2A. Patients were divided into the diagnostic subgroups of SZ-ND and SZ-D using the Schedule for Deficit Syndrome (SDS), and assessed clinically by the Positive and Negative Symptom Scale (PANSS). SNP8NRG241930 in NRG1 and rs1011313 in DTNBP1 were associated with SZ-ND (P ¼ 0.04 and 0.03, respectively). Polymorphisms in RGS4, G72/G30, and PIP5K2A were neither associated with SZ-ND nor with SZ-D. SNP8NRG241930 showed association with the PANSS cognitive and hostility/ excitability factors, rs1011313 with the negative factor and SDS total score, and rs10917670 in RGS4 was associated with the depression factor. Although these results replicate earlier findings about the genetic background of SZ-ND and SZ-D only partially, our data seem to confirm previously reported association of NRG1 with schizophrenia without prominent negative symptoms. It was possible to detect associations of small-to-medium effect size between the investigated candidate genes and symptom severity. Such studies have the potential to unravel the possible connection between genetic and clinical heterogeneity in schizophrenia. 2009 Wiley-Liss, Inc. Key words: schizophrenia; deficit syndrome; symptom severity; NRG1; DTNBP1 2009 Wiley-Liss, Inc. How to Cite this Article: Rethelyi JM, Bakker SC, Polgar P, Czobor P, Strengman E, Pasztor PI, Kahn RS, Bitter I. 2010. Association Study of NRG1, DTNBP1, RGS4, G72/G30, and PIP5K2A With Schizophrenia and Symptom Severity in a Hungarian Sample. Am J Med Genet Part B 153B:792–801. INTRODUCTION Family and twin studies have consistently demonstrated high heritability in schizophrenia [Sullivan et al., 2003], a devastating psychiatric disorder affecting 1% of the population. Schizophrenia is characterized by heterogeneous clinical symptoms, neurocognitive impairments, and decreased community functioning. Although the evidence for the genetic background of the disorder is well documented, the search for candidate genes in schizophrenia has resulted in conflicting findings. Initial reports based on the positional candidate gene approach have identified several promising genes with biologically plausible functions for the etiopathology of schizophrenia [Harrison and Weinberger, 2005]. However, many of the positive results found during the initial association Additional Supporting Information may be found in the online version of this article. *Correspondence to: Dr. Janos M. Rethelyi, Department of Psychiatry and Psychotherapy, Semmelweis University, 1083 Balassa u. 6., Budapest, Hungary. E-mail: firstname.lastname@example.org Published online 24 November 2009 in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/ajmg.b.31049 792 RETHELYI ET AL. studies could not be replicated in other populations. Methodological issues proposed to account for the low replicability of results were insufficient sample size and statistical power, ambiguously or too widely defined phenotypes, and initial chance findings as a result of multiple testing [Sullivan, 2007]. Meta-analyses have demonstrated small but significant associations for several of the candidate genes [Munafo et al., 2006, 2008; Allen et al., 2008]. The latest association studies and genome-wide association studies involving large samples have not confirmed the previously identified candidate genes in schizophrenia [Sanders et al., 2008; Sullivan et al., 2008], leading to skepticism about the usefulness of association studies in the search for the genetic background of schizophrenia. Neuregulin 1 (NRG1) and dysbindin (DTNBP1) were identified as schizophrenia candidate genes by fine-mapping of linkage regions 8p13 and 6p22, respectively [Stefansson et al., 2002; Straub et al., 2002]. The initially uncovered association of these genes with schizophrenia was replicated in some of the subsequent studies [Williams et al., 2005]. Other studies, however, yielded negative findings or positive findings implicating other markers and haplotypes within the same gene, suggesting locus heterogeneity [Harrison and Law, 2006]. Recent meta-analyses do not support the association of individual single-nucleotide polymorphisms (SNPs) in NRG1 with schizophrenia and show a considerable amount of between-study heterogeneity [Gardner et al., 2006]; but the haplotype-based analysis does reflect the association, although not for the originally described Iceland haplotype [Li et al., 2006; Munafo et al., 2006, 2008]. A recently performed high-resolution haplotype analysis of DTNBP1 [Mutsuddi et al., 2006] and the review of Guo et al.  point to major between-study differences in risk allele and haplotype frequencies. Similar to NRG1, locus heterogeneity and the effect of geographical differences were proposed as possible explanations. The SZGENE meta-analysis lends subtle support for the role of DTNBP1 as a candidate gene, pooled odds ratios for rs1011313[T] (P1325) demonstrated significant association with schizophrenia [Allen et al., 2008]. G72 and G30 are two overlapping genes, often referred to as D-amino-acid oxidase activator (DAOA) gene. Similar to the previously mentioned ones, it is a positional candidate gene that was localized during the fine-mapping of the 13q32-33 linkage region, where several SNPs showed association with schizophrenia [Chumakov et al., 2002]. Other association studies were inconclusive regarding the association of G72/G30 with schizophrenia, while meta-analytic data point again to locus heterogeneity [DeteraWadleigh and McMahon, 2006]. The regulator of G-protein signaling 4 (RGS4) gene was first implicated in schizophrenia based on expression studies that revealed decreased expression of the RGS4 gene in the prefrontal cortex of schizophrenic patients [Mirnics et al., 2001]. A subsequent association study uncovered a group of four SNPs that were associated with schizophrenia in three ethnically different samples [Chowdari et al., 2002]. Replication studies involving this candidate gene vary substantially, including studies showing the same association pattern or partial replication; one study failed to detect any association in a Han Chinese sample [Zhang et al., 2005]. The meta-analysis of Talkowski et al.  found a significant association for rs951436 (SNP4) in the analysis including all case–control samples (3,486 cases and 3,755 controls). 793 The same analysis detected no significant association with any of the SNPs or haplotypes in the family-based data set. The phosphatidylinositol-4-phosphate 5-kinase, type II, alpha (PIP5K2A) gene was identified as a schizophrenia candidate gene based on both functional and positional evidence. Its product is involved in the phosphoinositide signal transduction system, and it is also harbored by 10p12, a chromosomal region showing linkage to schizophrenia and bipolar disorder, although less consistently than other linkage regions [Stopkova et al., 2003]. Two association studies have reported several SNPs in PIP5K2A to be associated with schizophrenia [Stopkova et al., 2005; Schwab et al., 2006]. The studies of He et al.  and Saggers-Gray et al.  investigated the association of PIP5K2A in populations of non-European descent with negative results. Deficit schizophrenia (SZ-D) is a subgroup within schizophrenia that is characterized by enduring, idiopathic negative symptoms, including flattened affect, anhedonia, poverty of speech, curbing of interest, lack of sense of purpose, and diminished social drive [Carpenter et al., 1988]. These features are continuously present during periods of clinical stability and are not secondary, that is, not explainable by depression, anxiety, medication side effect, positive symptoms, substance abuse, or psychosocial deprivation. Since the first publication of the original concept a substantial body of clinical, pharmacologic, neuropsychological, and epidemiologic evidence has accumulated supporting the construct validity of the deficit syndrome as a pathophysiologically distinct subgroup within schizophrenia that affects about 15% of first-episode patients and 25–30% of patients with chronic schizophrenia [Kirkpatrick et al., 2001; Kirkpatrick and Galderisi, 2008]. Patients meeting criteria for the deficit syndrome initially suffer also from more severe longterm social and occupational disability [Tek et al., 2001] and have less chance for recovery [Strauss et al., 2008]. Genetic studies targeting the deficit syndrome as a hypothesized genetically unique subgroup within schizophrenia have provided further insight into this concept, although with mixed results. A genetic epidemiological study examining sibling pairs concordant for core schizophrenia revealed significant correlation of the deficit status in siblings [Ross et al., 2000]. Bakker et al. [2004, 2007] found association between non-deficit schizophrenia (SZ-ND) and previously reported risk haplotypes and SNPs of NRG1 and RGS4. No association was detectable with alleles or haplotypes of DTNBP1 and G72/G30 neither in the SZ-D nor in the SZ-ND groups. The PIP5K2A gene showed association with both diagnostic subtypes. In a study enrolling 86 cases and 50 healthy controls, Wonodi et al.  failed to show the association of the COMT Val158Met SNP with SZ-D. Only a few studies have investigated the association of candidate genes and symptom severity in schizophrenia. A DTNBP1 risk haplotype has been found to be associated with higher levels of negative symptomatology [DeRosse et al., 2006]. A study investigating the presence of the DTNBP1 C-A-T risk haplotype (rs2619539, rs3213207, and rs2619538) in patients of Irish origin found a significant association with the Positive and Negative Symptom Scale (PANSS) hostility/excitability factor, and a trend for negative symptoms [Corvin et al., 2008]. Another study conducted among Korean SZ patients showed a significant association between PANSS total and positive scores and another DTNBP1 794 haplotype (rs3213207 and rs1011313) incorporating one common SNP with the previously mentioned C-A-T risk haplotype [Pae et al., 2008]. Campbell et al. [2008a,b] reported the association of more severe baseline PANSS total scores in patients participating in the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) with the markers rs2661319 (SNPRGS4-18) and rs2842030 in RGS4 and rs10799902 in RGS5. The rs3918346 risk allele in the D-amino-oxidase gene has been reported to show an association with the depression/anxiety PANSS score [Corvin et al., 2007]. Finally, Yue et al.  found the DAOA gene AAG risk haplotype (rs2391191, rs947267, and rs778294) to be associated with the PANSS negative, cognitive, and depression factors in Han Chinese patients. Overall, the reported associations between symptom severity and risk alleles or haplotypes show a substantial variation among studies and the effect sizes were modest (values typically fell in the small-to-medium range). Due to the variation of investigated candidate genes the results are difficult to compare across studies. The principal objective of the present study was to investigate the association of previously reported candidate genes for schizophrenia with SZ-ND and SZ-D in an ethnically homogeneous sample of Hungarian schizophrenic patients, a population scarcely investigated previously in psychiatric genetic studies. Our primary goal was the replication of the results of a prior Dutch study. The original studies of Bakker et al. [2004, 2007] investigated the association of candidate genes with diagnostic subgroups of schizophrenia that represent the lack or the concentration of negative symptomatology. The underlying hypothesis for such an approach is that narrowing the broader phenotype of schizophrenia, that is, decreasing phenotypic heterogeneity, increases the chances of detecting a genetic association. In order to account for the phenotypic variation and heterogeneity further, we complemented the above categorical approach (i.e., contrasting diagnostic subgroups) with a dimensional one. In particular, we explored the association of the investigated candidate genes with specific symptom clusters and severity in the patient group. Based on previous results [Bakker et al., 2004, 2007; DeRosse et al., 2006] we hypothesized the association of NRG1 and RGS4 with SZ-ND, and the association of DTNBP1 with SZ-D. In addition, in light of the aforementioned findings we expected that higher level of negative symptomatology would occur in association with the DTNBP1 risk alleles, and higher positive symptomatology in association with the NRG1 and RGS4 risk alleles. METHODS Sample Collection, Clinical, and Psychopathological Assessment Inpatients and outpatients with a DSM-IV diagnosis [American Psychiatric Association, 1994] of schizophrenia (n ¼ 284) were enrolled in the study after written informed consent from two major psychiatric centers in Budapest, Hungary (Department of Psychiatry and Psychotherapy, Semmelweis University and Szent Janos Hospital, Psychiatry Unit). Criteria for exclusion were severely disorganized behavior that prevented patient cooperation AMERICAN JOURNAL OF MEDICAL GENETICS PART B and precluded testing, and severe comorbidity, such as neurological disorders, head trauma, mental retardation, or substance abuse. All patients underwent clinical testing; a subsample of 266 patients participated also in neuropsychological testing (results not presented in this study). The DSM-IV diagnosis for schizophrenia excluding schizoaffective disorder was validated using the MINI 5.0 Neuropsychiatric Interview [Balazs et al., 1998]. The Hungarian version of the Schedule for the Deficit Syndrome (SDS) was obtained for all patients [Kirkpatrick et al., 1989], out of whom 154 met the criteria for the deficit syndrome. The SDS assesses enduring negative symptoms (flattened affect, anhedonia, poverty of speech, curbing of interest, lack of sense of purpose, diminished social drive) based on the present condition and the history of the patient. Following the original instructions, patients were assigned to the SZ-D group if at least two of these symptoms were present in an enduring manner, and all other underlying causes could be ruled out. The PANSS [Kay et al., 1987] was administered by trained staff members. Healthy controls (n ¼ 238) were recruited from the employees of Semmelweis University (Budapest, Hungary) and outpatients of the Department of Internal Medicine after screening for psychiatric disorders. The study was approved by the Hungarian National Scientific and Ethical Committee (ETT TUKEB) and the Semmelweis University Institutional Ethical Board. DNA samples of 280 schizophrenic (127 non-deficit and 153 deficit) patients and 230 controls were available; these subjects represented the target population for this study. Basic demographic and clinical data describing the two diagnostic subgroups and the control group are summarized in Table I. Chi-square and General Linear Model (GLM) analysis showed differences between the groups in terms of age, level of education, duration of illness, and psychopathology scores. Marker Selection The selection of SNPs was based on the marker selection and results of prior Dutch reports of Bakker et al. [2004, 2007], and the existing literature on candidate genes (mentioned earlier). In the case of SNPs forming haplotype blocks in the Dutch sample, one of them was omitted (rs760761, rs2619522, rs3916967, and rs1421292). rs746203 (hCV9591220) was omitted because of the lack of any association in the original sample. Additional SNPs of potential importance (SNP8NRG221132, SNP8NRG241930, SNP8NRG243177, and rs909706) were added to the set of markers in accordance with recent studies, in order to increase compatibility with these results [Burdick et al., 2007; Stefanis et al., 2007]. Genotyping Genotyping was performed based on the TaqMan 50 -exonuclease allelic discrimination assay, using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA). Polymerase chain reaction (PCR) for each SNP was performed with 5 ng of DNA, 2.0 ml of ABI Taqman Universal Mastermix, 1.9 ml of water, and 0.1 ml of 20 SNP assay (ABI). The 5 ml reactions were performed in a 384-well plate (ABI). The DNA samples were PCR amplified using the following conditions: 2 min hold at 50 C, 10 min denaturation step at 95 C, 50 cycles of 92 C for 15 sec, and RETHELYI ET AL. 795 TABLE I. Clinical and Demographic Characteristics of the SZ Subgroups and the Control Group Age (SD) Sex (M:F, %) Level of education (E:S:H ratio, %)b Education (years, SD)c Age at onset (SD) Duration of illness (SD) No. of hospitalizations (mean, SD) Inpatient:outpatient ratio (%) SDS total score (SD) PANSS total (SD) PANSS positive (SD) PANSS negative (SD) PANSS general (SD) SZ-ND (n ¼ 127) SZ-D (n ¼ 153) Control (n ¼ 230) F/x2a P-value 36.2 (11.2) 38.8 (12.0) 39.9 (15.0) 3.18 0.04 59:68 (46%:54%) 71:82 (46%:54%) 98:132 (42%:58%) 1.06 0.58 38:54:35 (30%:43%:27%) 56:81:16 (36%:53%:11%) 52:115:63 (23%:50%:27%) 23.9 <0.0001 13.4 (3.1) 11.9 (2.8) 13.9 (3.2) 12.49 <0.0001 28.4 (9.5) 28.8 (9.8) — 0.12 0.73 7.5 (8.2) 10.1 (9.5) — 5.34 0.021 5.3 (4.9) 6.9 (6.0) — 5.63 0.02 81:46 (64%:36%) 118:35 (77%:23%) — 7.94 0.02 10.7 (2.0) 16.6 (2.9) — 381.10 <0.0001 72.0 (14.7) 83.5 (14.0) — 44.90 <0.0001 17.9 (4.9) 17.6 (4.9) — 0.14 0.71 17.1 (4.6) 22.4 (4.3) — 100.10 <0.0001 37.0 (8.4) 43.4 (8.1) — 42.38 <0.0001 a Results of GLM and c2 statistics and corresponding levels of significance. For completed education, E ¼ elementary, S ¼ secondary, H ¼ higher education. c Available for 158 patients and 205 controls. b 60 C for 1 min. The plates were scanned utilizing the Allelic Discrimination End-Point Analysis on the ABI Prism 7900HT Sequence Detection System. The allelic discrimination data were analyzed by the AutoCall algorithm of the SDS v2.1 software (ABI). To ensure reliability, all PCR plates contained 16 blind duplicate samples. Statistical Methods The Statistical Analysis System for Windows (version 9.1; SAS Institute, Cary, NC) was used for statistical analyses. Genotype frequencies in controls were checked for Hardy–Weinberg equilibrium (HWE) using a SAS macro calculating the 2 ln Q and chisquare test statistics for the hypothesis of HWE and the corresponding observed levels of significance. Allele frequencies were calculated and cross-tabulated for the deficit and non-deficit group, the total schizophrenia group, and controls. Association of alleles with disease groups was tested using the Generalized Linear Model (GENMOD) analysis, where alleles of interest served as a dependent variable (in separate analyses), and disease group was used as an independent variable; the likelihood ratio statistics determined from the model tested the association. Statistical significance was determined for the whole model, and for the subgroup comparisons, using model contrasts. All P-values are two-tailed and represent comparison-wise type 1 error rates, with no adjustment for multiple testing. Major haplotypes were calculated, and their association with schizophrenia and the subgroups was analyzed using the Unphased software (version 3.0.12) [Dudbridge, 2008]. Linkage disequilibrium (LD) analyses and comparison of LD results with HapMap data [The International HapMap Consortium, 2005] were performed using the Haploview 4.1 software [Barrett et al., 2005]. The association of symptom severity with SNPs was tested using the GLM analysis. In the GLM model, genotypes (in separate analyses) were the independent variables (risk allele carriers vs. non-carriers). The PANSS total score, the five PANSS factors, identified by Varimax rotation principal component analysis, and the SDS total score were applied as dependent variables; age and sex were added as covariates to the analyses. The identified PANSS factors (positive, negative, hostility/excitability, cognitive, depression) were in accordance with the literature [Lindenmayer et al., 1995] both in terms of item composition and explained variance. The item composition of the five factors is listed in Supplementary Material Table I. Similar to previous reports, the total variance explained by the five factors was 57.3%, indicating the assay sensitivity of clinical testing. In the GLM analyses testing the association of PANSS positive and negative factors the opposite factors (positive or negative, depending on analysis) were included as covariates to distinguish primary and secondary negative symptoms. Indicators of symptom severity were compared between genotypes by effect size estimation [Cohen, 1988]. By definition, healthy controls were omitted from the analyses of symptom severity. The analysis of continuous variables, rather than categorical variables, holds the advantage of an increased statistical power to as much as 50% [Selvin, 2004]. RESULTS Genotyping Results: General Information and Intermarker Linkage Disequilibrium (LD) The investigated candidate genes were genotyped altogether for 14 SNPs. The rate of successful genotyping was above 97% for all SNPs, except for rs2619528 in DTNBP1, where 90.2% of the samples were genotyped. All markers except for rs2619539 were in HWE, defining the P-value for HWE at the level of 0.001 (Supplementary Material Table II). This marker was omitted from further analyses. Patterns of LD between the SNPs for the whole sample are shown in Figure 1. In NRG1 SNP8NRG221533 and SNP8NRG243177 composed a haplotype block (D0 ¼ 0.90, r2 ¼ 0.70). SNPs rs3213207 796 AMERICAN JOURNAL OF MEDICAL GENETICS PART B FIG. 1. NRG1 and DTNBP1 intermarker linkage disequilibrium measured with D0 . D0 values were calculated from the genotyping data obtained from the entire sample, using Haploview. and rs2619528 in DTNBP1 (D0 ¼ 0.951, r2 ¼ 0.549) and rs10917670 and rs2661319 in RGS4 (D0 ¼ 0.879, r2 ¼ 0.546) were in close LD. There was hardly any LD between rs2391191 and rs3918342 in G72/G30. SZ-D subgroups, and the entire schizophrenia sample (SZ). Resulting P-values of likelihood tests indicate statistical significance comparing the control sample and the total schizophrenia sample, and contrast between the control sample and both the SZ-ND and SZ-D subgroups. For NRG1, the SNP8NRG241930[G] marker was associated with SZ-ND (P ¼ 0.04), but not with SZ-D, nor with the total SZ sample. No additional SNP in NRG1 showed an association with SZ, or the subgroups. The two-marker haplotypes NRG221132–241930 and NRG241930–243177 both were associated with SZ-ND as single markers (P ¼ 0.03 and 0.04, respectively), although the overall association of the haplotypes did not reach the level of statistical significance. For DTNBP1, rs1011313[C] showed association with the SZ-ND group (P ¼ 0.03) but not with the SZ-D group or the entire SZ sample. The association approached significance for rs909706 in the total SZ sample (P ¼ 0.069). Analysis of two-marker haplotypes yielded a significant association between rs909707–rs3213207 and the total SZ sample (P ¼ 0.05). Neither of the markers nor the two-marker haplotype in RGS4 was associated with SZ-ND or SZ-D, or the pooled SZ sample. For G72/G30, we detected a marginally significant association between rs2391191 and SZ-D (P ¼ 0.07), otherwise there was no association between rs3918342 or the haplotype with any of the groups. For PIP5K2A, the tested marker rs10828317 was not associated with any of the disease subgroups. Association of the Investigated SNPs and Haplotypes With Schizophrenia, Non-Deficit and Deficit-Schizophrenia Association of Investigated Genotypes With Symptom Severity Table II summarizes allele frequencies of all selected markers, and major haplotype frequencies in the control sample, the SZ-ND and Results of the GLM analyses showed that the association of psychopathological data (PANSS factor and total scores, and the TABLE II. Association of Investigated SNPs and Haplotypes With Schizophrenia, SZ-ND and SZ-D Schizophrenia Gene NRG1 DTNBP1 Haplo RGS4 G72/G30 PIP5K2A Marker/dbSNP NRG221132 NRG221533 NRG241930 NRG243177 Haplo rs1011313 rs909706 rs3213207 rs2619528 rs909706–rs3213207 rs10917670 rs2661319 Haplo rs2391191 rs3918342 Haplo rs10828317 Allele C C G C Cont. 87.8 35.6 65.9 60.5 SZ-ND 89.7 39.0 73.6 57.1 C C C C CG A C 88.6 54.6 13.4 81.1 41.5 42.8 47.1 93.3 60.3 11.4 81.5 48.8 43.3 48.4 A C 41.1 43.8 40.8 44.6 C 29.1 25.2 P 0.43 0.39 0.04 0.38 n.s. 0.03 0.15 0.43 0.90 0.17 0.88 0.74 n.s. 0.93 0.83 n.s. 0.27 SZ-D 87.8 36.1 66.5 63.2 90.8 60.6 10.2 82.1 49.3 48.3 44.0 34.9 49.3 27.3 Cont., control group (n ¼ 230); SZ-ND, non-deficit schizophrenia (n ¼ 127); SZ-D, deficit-schizophrenia (n ¼ 153); SZ: all schizophrenics (n ¼ 280). P 0.98 0.90 0.87 0.45 n.s. 0.32 0.11 0.16 0.74 0.09 0.11 0.40 n.s. 0.07 0.16 n.s. 0.60 SZ 88.67 37.4 69.7 60.4 91.9 60.5 10.8 81.8 49.1 46.1 46.0 37.6 47.3 26.4 P 0.66 0.58 0.21 0.97 n.s. 0.08 0.069 0.19 0.77 0.05 0.27 0.72 n.s. 0.24 0.29 n.s. 0.34 RETHELYI ET AL. SDS total score) with genetic variables approached statistical significance. Findings are summarized below in this section in order to characterize the effect sizes for the pertinent associations for subsequent meta-analytic investigations. For NRG1, SNPNRG241930[G] was associated with the cognitive (d ¼ 0.48) and hostility/excitability factor (d ¼ 0.37) scores, and the PANSS total score (d ¼ 0.21). Notably, the association was opposite to the direction expected by meta-analytic data. For DTNBP1, rs1011313[T] was associated marginally with the negative factor score (d ¼ 0.28) and the SDS total score (d ¼ 0.21), while rs909706[T] (d ¼ 0.23), rs3213207[T] (d ¼ 0.26), and rs2619528[T] (d ¼ 0.23) showed association with the positive factor score. rs10917670[G] in RGS4 showed a statistically significant association (d ¼ 0.31) with the PANSS depression factor score. The observed associations resulted in low (d ¼ 0.2–0.4) or medium (d ¼ 0.4–0.5) effect sizes. Results on symptom severity are summarized in Table III. DISCUSSION The primary objective of this study was to replicate the results of the original association study of Bakker et al. [2004, 2007], focusing on negative symptomatology, in an ethnically similar, however, geographically distinct population of schizophrenic patients. The original study took place in the Netherlands and was limited to patients with Dutch, Caucasian origin. Patients fulfilling criteria for SZ-D were overrepresented. The replication phase followed the original study in most aspects, that is, Hungarian schizophrenic patients, exclusively with Caucasian ethnic background were enrolled in the study, and the proportion of SZ-D patients was similar to that in the original study. The large proportion of SZ-D patients in the Hungarian sample is explicable by the patient characteristics at the study sites, two inpatient units involved in the treatment of chronic, severely ill patients. The number of healthy controls was less in the replication phase; however, it is important to note that the control group in the Hungarian study was composed by individually recruited healthy subjects, instead of bloodbank controls. The SDS was used for clinical assessment in both studies. Validation of the clinical diagnosis of patients was done with different diagnostic interviews, the Comprehensive Assessment of Symptoms and History (CASH) [Andreasen et al., 1992] in the prior study, and the MINI 5.0 Neuropsychiatric Interview in the replication study; however, both instruments have been shown to be reliable. The use of the SDS to differentiate patients according to the concentration of negative symptomatology was the core concept of both studies. The common approach was based on the assumption that genetic variation may be decreased by the reduction of phenotypic variation, by investigating homogenous subgroups, and thus making the detection of genetic association more probable. The importance of this replication study is underscored by the fact that Hungarian schizophrenia patients have only been studied in a few psychiatric genetic studies [Schwab et al., 2003; Szekeres et al., 2004]. Accumulating association study findings in schizophrenia point to the importance of locus heterogeneity, a phenomenon that can be better understood by replicating studies in different regions and populations [Straub and Weinberger, 2006]. 797 The original reports of Bakker et al. found SNP8NRG221533 in NRG1 and rs10917670 in RGS4 to be associated with SZ-ND, furthermore rs10828317 in PIP5K2A to be associated with both SZ-ND and SZ-D [Bakker et al., 2004, 2007]. In the Hungarian sample, we detected association of NRG1 similarly only with SZ-ND (P ¼ 0.04), although with another variant, SNP8NRG241930[G]. This marker has been shown to be associated with schizophrenia in earlier studies as well [Petryshen et al., 2005], although the pooled odds ratio estimate of all published association studies was not statistically significant (OR ¼ 0.96; 95% CI: 0.88–1.04); it approached statistical significance if the meta-analysis was restricted to only the Caucasian studies (OR ¼ 0.93; 95% CI: 0.85–1.03) [Allen et al., 2008]. As in the Dutch sample, allele frequencies in SZ-D were almost indistinguishable from those of controls. Several lines of evidence indicate that NRG1 is involved in psychosis phenotypes with relative preservation of affect and better prognosis, such as SZ-ND, bipolar disorder with mood-incongruent psychotic features, or schizotypal personality traits [Green et al., 2005; Lin et al., 2005]. Both the original study and the replication are consistent with these earlier findings. It is conceivable that across populations different SNPs are associated with a nearby genetic variant that is associated with psychotic traits without prominent negative symptoms. The distribution of allelic frequencies across groups and the direction of effect sizes for SNP8NRG221533, the common NRG1 marker in the two studies were opposite (Supplementary Material Table III). Marker rs1011313[C] in DTNBP1 was associated with SZ-ND, and the two-marker haplotype composed by rs909706 and rs3213207 was associated with SZ. The SZGene meta-analysis demonstrated significant pooled ORs for rs1011313[T] in Caucasian samples (OR ¼ 1.23; 95% CI: 1.07–1.40), thus the distribution of allelic frequencies in the majority of previous studies was opposite [Allen et al., 2008]. A possible explanation for the contradictory result is that the C allele of rs1011313 is a protective allele that is associated with reduced risk for negative symptoms and SZ-D in schizophrenia patients. We found no association of RGS4 and PIP5K2A markers with any of the subgroups, and an association approaching significance for rs2391191 in G72/G30 with SZ-D. The results observed in the Hungarian study are at variance with the original study, where RGS4 was associated with SZ-ND and PIP5K2A with both groups. It should be noted, however, that the observed allelic variations fall in the range of previous meta-analytic data and are similar in direction in the two samples (Supplementary Material Table III). In addition, the meta-analysis of the combined (the Netherlands and Hungarian) sample resulted in a low, however, statistically significant effect size for rs10828317[T] in PIP5K2A (h ¼ 0.14; 95% CI: 0.25 to 0.04). Furthermore, we only studied a limited number of SNPs that showed association in previous studies. Association of other SNPs in the Hungarian population can therefore not be ruled out yet. It is important to note that while our investigation might have been underpowered to detect associations with a small effect size, neither study could identify a specific association of the deficit syndrome with the investigated candidate genes. It is conceivable that SZ-D is associated with other markers, or other sources of genetic variation, as suggested by recent research about the role of copy number variations in schizophrenia [Vrijenhoek et al., 2008]. PANSS negative factor PANSS hostility/excitability factor PANSS cognitive factor Marker/dbSNP 221132[C]a 221533[C] 241930[G] 243177[T] rs1011313[T] rs909706[T] rs3213207[T]a rs2619528[T] rs10917670[G] rs2661319[C] rs2391191[A] rs3918342[T] rs10828317[T] Risk all. carrier (%) 79.5 62.3 89.6 63.8 15.5 62.5 78.8 34.7 79.2 68.6 61.4 76.6 90.9 Risk all. carrier 10.5 10.4 10.5 10.5 10.5 10.5 10.5 10.2 10.3 10.5 10.6 10.4 10.5 Risk allele non-carrier 10.3 10.6 10.4 10.5 10.5 10.5 10.4 10.7 11.3 10.6 10.3 10.7 10.8 Effect size (Cohen d) 0.06 0.06 0.03 0.00 0.00 0.00 0.03 0.16 0.31 0.03 0.09 0.09 0.09 PANSS depression factor Comparisons were calculated between risk allele homozygotes and other genotypes due to low case numbers. a PIP5K2A G72/G30 RGS4 DTNBP1 Gene NRG1 Gene NRG1 Risk all. carrier 77.9 78.1 78.0 78.0 79.9 78.4 78.4 77.6 78.4 78.3 78.6 77.6 79.3 Risk allele non-carrier 80.0 78.9 81.2 78.7 78.1 78.0 77.8 78.5 79.5 78.5 78.0 79.8 79.2 Effect size (Cohen d) 0.14 0.05 0.21 0.05 0.12 0.03 0.04 0.06 0.07 0.01 0.04 0.14 0.01 PANSS total score Risk all. carrier 13.7 13.9 13.9 13.7 14.6 13.8 13.9 13.8 14.0. 13.8 13.9 13.7 13.9 Risk allele non-carrier 14.0 14.2 14.5 14.1 13.8 14.1 13.8 13.8 13.9 14.3 14.0 14.5 14.3 SDS total score Effect size (Cohen d) 0.17 0.06 0.16 0.09 0.21 0.08 0.03 0.00 0.03 0.13 0.05 0.19 0.10 Risk all. Risk all. Risk allele Effect size Risk all. Risk allele Effect size Risk all. Risk allele Effect size Risk all. Risk allele Effect size Marker/dbSNP carrier (%) carrier non-carrier (Cohen d) carrier non-carrier (Cohen d) carrier non-carrier (Cohen d) carrier non-carrier (Cohen d) 221132[C]a 79.5 14.7 14.7 0.00 21.9 22.9 0.16 10.1 10.5 0.11 15.6 16.3 0.17 221533[C] 62.3 14.8 14.8 0.00 22.1 22.1 0.00 10.2 10.2 0.00 15.6 16.1 0.12 241930[G] 89.6 14.7 14.7 0.00 22.1 21.8 0.05 10.1 11.4 0.37 15.6 17.5 0.48 243177[T] 63.8 14.7 14.7 0.00 22.1 22.1 0.00 10.2 10.2 0.00 15.5 16.2 0.17 DTNBP1 rs1011313[T] 15.5 14.1 14.8 0.16 23.5 21.8 0.28 10.4 10.2 0.06 16.0 15.7 0.07 rs909706[T] 62.5 15.1 14.1 0.23 21.7 22.7 0.16 10.4 9.8 0.17 15.8 15.7 0.02 78.8 14.9 13.8 0.26 22.0 22.0 0.00 10.2 10.0 0.06 15.6 16.4 0.20 rs3213207[T]a rs2619528[T] 34.7 14.0 15.0 0.23 22.3 21.9 0.07 9.7 10.4 0.20 16.3 15.5 0.20 RGS4 rs10917670[G] 79.2 14.7 14.9 0.05 21.9 22.2 0.05 10.3 10.4 0.03 15.9 15.9 0.00 rs2661319[C] 68.6 14.8 14.6 0.05 21.8 22.5 0.11 10.3 10.1 0.06 15.8 15.7 0.02 G72/G30 rs2391191[A] 61.4 14.8 14.6 0.05 21.8 22.3 0.08 10.2 10.2 0.00 15.9 15.7 0.05 rs3918342[T] 76.6 14.6 14.9 0.07 21.9 22.5 0.10 10.0 10.6 0.17 15.6 15.9 0.07 PIP5K2A rs10828317[T] 90.9 14.7 15.1 0.09 22.1 22.3 0.03 10.2 9.9 0.09 15.7 15.9 0.05 PANSS positive factor TABLE III. Association of the Investigated Genotypes With PANSS Factor and Total Scores and the SDS Total Score 798 AMERICAN JOURNAL OF MEDICAL GENETICS PART B RETHELYI ET AL. Caution must be exercised while interpreting the above findings. As mentioned earlier, the results are not adjusted for multiple comparisons, neither for the number of SNPs in one candidate gene nor for all investigated associations. The results would not withstand this correction, and therefore can only be interpreted in the context of meta-analytic data. However, based on our a priori calculations our sample has the power to detect associations of the magnitude that were observed in the Netherlands sample. Practical issues explaining the low sample size were that we restricted the recruitment of patients to two clinical centers and we sought to have a sample where the proportion of deficit patients is higher. Using factor analysis we identified five PANSS factors that served as indicators of symptom clusters and severity. It should be emphasized that the factor structure of the PANSS yielded by our factor analyses was consistent with earlier results and pointed to high assay sensitivity of the clinical assessment. Testing the relationship of the investigated SNPs with PANSS total and factor scores revealed several associations. The positive factor was associated with four markers in DTNBP1. Negative factor scores and the SDS total score were significantly higher in carriers of the rs1011313 risk allele. The risk haplotype of DTNBP1 has been shown to be associated with higher levels of negative symptomatology and worse cognitive decline in schizophrenia [Funke et al., 2004; DeRosse et al., 2006; Burdick et al., 2007]. Moreover, Pae et al.  observed an association of a two-marker haplotype containing rs1011313 with total PANSS scores. Unfortunately, the originally described DTNBP1 risk haplotype cannot be investigated in our sample, since data on rs1018381 (P1578), the marker most significantly associated with SZ in the original sample, are not available. The association of the cognitive and the hostility/excitability factors, and the PANSS total score approached significance with SNP8NRG241930[G], the marker that was associated significantly with the SZ-ND group. Notably, the sign of association for the cognitive factor was negative, which is consistent with the finding that SZ-D patients display a higher level of cognitive impairment compared to SZ-ND patients. It is noteworthy that the depression factor was associated significantly with rs10917670[G] in RGS4, which showed no association with the SZ sample or the subgroups. Although rs10917670 was not included in the study of Campbell et al. [2008a], in that sample rs2661319 (SNP18) was significantly associated with the PANSS total score. Overall, the magnitude of the observed associations was similar to those reported in earlier studies, representing effect sizes in the small-to-medium rage. The interpretation of these findings is difficult, since similar to results of prior investigations, the associations that we found were modest, and in some cases were at variance with comparable previous studies. It should be noted that despite the fact that the factor-analytic technique identifies basic underlying symptoms of schizophrenia and allows for an efficient compression of psychopathological data, PANSS factors are still likely to reflect complex phenotypes, incorporating both trait- and state-related symptoms. Thus, prospective studies focusing on temporally stable symptom clusters may be needed to unravel the connection between symptomatological and genetic heterogeneity in schizophrenia. A recently published study successfully identified two SNPs in genes not implicated earlier for schizophrenia that were 799 associated with lifetime positive and disorganized symptom severity [DeRosse et al., 2008]. To summarize, our study demonstrated the feasibility of defining particular patient groups with specific characteristics for genetic association studies. Our data are suggestive of an association of NRG1 with SZ-ND and lend support to the idea that the power of genetic studies may be increased by analyzing disease subtypes. In addition, we tested the association of psychopathological clusters with the investigated candidate genes, resulting in several associations of interest. These novel approaches have the potential of clarifying the connection between multiple genetic factors and clinical heterogeneity in schizophrenia and might ultimately lead to the identification of further candidate genes, or other sources of genetic variation that play a role in the etiopathology of schizophrenia. ACKNOWLEDGMENTS The authors are indebted to Vikt oria Szab o and Adrien St€ undl, who helped in preparing DNA samples; furthermore, to Beatrix Mersich, Krisztina Magyar, Agnes Fabian, and Krisztina Jarvas for their assistance in clinical data collection. We also wish to thank the R help of Csilla Boly os, Marta Farkas, Zita Murai, Eva ozsav€ olgyi, Tamas Kurimay, Henrik Horvath, and Gabor Gazdag in recruiting patients and healthy controls to the study. We are grateful to Roel Ophoff for his advice and support at all stages of the study. REFERENCES Allen NC, Bagade S, McQueen MB, Ioannidis JP, Kavvoura FK, Khoury MJ, Tanzi RE, Bertram L. 2008. Systematic meta-analyses and field synopsis of genetic association studies in schizophrenia: The SzGene database. Nat Genet 40:827–834. American Psychiatric Association. 1994. Diagnostic and statistical manual of mental disorders: DSM-IV. Washington, DC: American Psychiatric Association Press. Andreasen NC, Flaum M, Arndt S. 1992. The Comprehensive Assessment of Symptoms and History (CASH). An instrument for assessing diagnosis and psychopathology. Arch Gen Psychiatry 49:615–623. Bakker SC, Hoogendoorn ML, Selten JP, Verduijn W, Pearson PL, Sinke RJ, Kahn RS. 2004. Neuregulin 1: Genetic support for schizophrenia subtypes. Mol Psychiatry 9:1061–1063. Bakker SC, Hoogendoorn ML, Hendriks J, Verzijlbergen K, Caron S, Verduijn W, Selten JP, Pearson PL, Kahn RS, Sinke RJ. 2007. The PIP5K2A and RGS4 genes are differentially associated with deficit and non-deficit schizophrenia. Genes Brain Behav 6:113–119. Balazs J, Bitter I, Hideg K, Vitrai J. 1998. The Hungarian version of the M.I.N.I. and the M.I.N.I. Plus (in Hungarian). Psychiatr Hung 13: 160–168. Barrett JC, Fry B, Maller J, Daly MJ. 2005. Haploview: Analysis and visualization of LD and haplotype maps. Bioinformatics 21:263–265. Burdick KE, Goldberg TE, Funke B, Bates JA, Lencz T, Kucherlapati R, Malhotra AK. 2007. DTNBP1 genotype influences cognitive decline in schizophrenia. Schizophr Res 89:169–172. Campbell DB, Ebert PJ, Skelly T, Stroup TS, Lieberman J, Levitt P, Sullivan PF. 2008a. Ethnic stratification of the association of RGS4 variants with antipsychotic treatment response in schizophrenia. Biol Psychiatry 63:32–41. 800 Campbell DB, Lange LA, Skelly T, Lieberman J, Levitt P, Sullivan PF. 2008b. Association of RGS2 and RGS5 variants with schizophrenia symptom severity. Schizophr Res 101:67–75. Carpenter WT Jr, Heinrichs DW, Wagman AM. 1988. Deficit and nondeficit forms of schizophrenia: The concept. Am J Psychiatry 145: 578–583. Chowdari KV, Mirnics K, Semwal P, Wood J, Lawrence E, Bhatia T, Deshpande SN, Thelma BK, Ferrell RE, Middleton FA, et al. 2002. Association and linkage analyses of RGS4 polymorphisms in schizophrenia. Hum Mol Genet 11:1373–1380. Chumakov I, Blumenfeld M, Guerassimenko O, Cavarec L, Palicio M, Abderrahim H, Bougueleret L, Barry C, Tanaka H, La Rosa P, et al. 2002. Genetic and physiological data implicating the new human gene G72 and the gene for D-amino acid oxidase in schizophrenia. Proc Natl Acad Sci USA 99:13675–13680. Cohen J. 1988. Statistical power analysis for the behavioral sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Corvin A, Donohoe G, McGhee K, Murphy K, Kenny N, Schwaiger S, Nangle JM, Morris D, Gill M. 2007. D-amino acid oxidase (DAO) genotype and mood symptomatology in schizophrenia. Neurosci Lett 426:97–100. Corvin A, Donohoe G, Nangle JM, Schwaiger S, Morris D, Gill M. 2008. A dysbindin risk haplotype associated with less severe manic-type symptoms in psychosis. Neurosci Lett 431:146–149. DeRosse P, Funke B, Burdick KE, Lencz T, Ekholm JM, Kane JM, Kucherlapati R, Malhotra AK. 2006. Dysbindin genotype and negative symptoms in schizophrenia. Am J Psychiatry 163:532–534. DeRosse P, Lencz T, Burdick KE, Siris SG, Kane JM, Malhotra AK. 2008. The genetics of symptom-based phenotypes: Toward a molecular classification of schizophrenia. Schizophr Bull 34:1047–1053. Detera-Wadleigh SD, McMahon FJ. 2006. G72/G30 in schizophrenia and bipolar disorder: Review and meta-analysis. Biol Psychiatry 60:106–114. Dudbridge F. 2008. Likelihood-based association analysis for nuclear families and unrelated subjects with missing genotype data. Hum Hered 66:87–98. Funke B, Finn CT, Plocik AM, Lake S, DeRosse P, Kane JM, Kucherlapati R, Malhotra AK. 2004. Association of the DTNBP1 locus with schizophrenia in a U.S. population. Am J Hum Genet 75:891–898. Gardner M, Gonzalez-Neira A, Lao O, Calafell F, Bertranpetit J, Comas D. 2006. Extreme population differences across neuregulin 1 gene, with implications for association studies. Mol Psychiatry 11:66–75. Green EK, Raybould R, Macgregor S, Gordon-Smith K, Heron J, Hyde S, Grozeva D, Hamshere M, Williams N, Owen MJ, et al. 2005. Operation of the schizophrenia susceptibility gene, neuregulin 1, across traditional diagnostic boundaries to increase risk for bipolar disorder. Arch Gen Psychiatry 62:642–648. Guo AY, Sun J, Riley BP, Thiselton DL, Kendler KS, Zhao Z. 2009. The dystrobrevin-binding protein 1 gene: Features and networks. Mol Psychiatry 14:18–29. Harrison PJ, Law AJ. 2006. Neuregulin 1 and schizophrenia: Genetics, gene expression, and neurobiology. Biol Psychiatry 60:132–140. Harrison PJ, Weinberger DR. 2005. Schizophrenia genes, gene expression, and neuropathology: On the matter of their convergence. Mol Psychiatry 10:40–68. He Z, Li Z, Shi Y, Tang W, Huang K, Ma G, Zhou J, Meng J, Li H, Feng G, et al. 2007. The PIP5K2A gene and schizophrenia in the Chinese population—A case–control study. Schizophr Res 94:359–365. Kay SR, Fiszbein A, Opler LA. 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull 13:261–276. AMERICAN JOURNAL OF MEDICAL GENETICS PART B Kirkpatrick B, Galderisi S. 2008. Deficit schizophrenia: An update. World Psychiatry 7:143–147. Kirkpatrick B, Buchanan RW, McKenney PD, Alphs LD, Carpenter WT Jr. 1989. The Schedule for the Deficit syndrome: An instrument for research in schizophrenia. Psychiatry Res 30:119–123. Kirkpatrick B, Buchanan RW, Ross DE, Carpenter WT Jr. 2001. A separate disease within the syndrome of schizophrenia. Arch Gen Psychiatry 58:165–171. Li D, Collier DA, He L. 2006. Meta-analysis shows strong positive association of the neuregulin 1 (NRG1) gene with schizophrenia. Hum Mol Genet 15:1995–2002. Lin HF, Liu YL, Liu CM, Hung SI, Hwu HG, Chen WJ. 2005. Neuregulin 1 gene and variations in perceptual aberration of schizotypal personality in adolescents. Psychol Med 35:1589–1598. Lindenmayer JP, Grochowski S, Hyman RB. 1995. Five factor model of schizophrenia: Replication across samples. Schizophr Res 14:229–234. Mirnics K, Middleton FA, Stanwood GD, Lewis DA, Levitt P. 2001. Diseasespecific changes in regulator of G-protein signaling 4 (RGS4) expression in schizophrenia. Mol Psychiatry 6:293–301. Munafo MR, Thiselton DL, Clark TG, Flint J. 2006. Association of the NRG1 gene and schizophrenia: A meta-analysis. Mol Psychiatry 11: 539–546. Munafo MR, Attwood AS, Flint J. 2008. Neuregulin 1 genotype and schizophrenia. Schizophr Bull 34:9–12. Mutsuddi M, Morris DW, Waggoner SG, Daly MJ, Scolnick EM, Sklar P. 2006. Analysis of high-resolution HapMap of DTNBP1 (dysbindin) suggests no consistency between reported common variant associations and schizophrenia. Am J Hum Genet 79:903–909. Pae CU, Drago A, Kim JJ, Patkar AA, Jun TY, Lee C, Mandelli L, De Ronchi D, Paik IH, Serretti A. 2008. DTNBP1 haplotype influences baseline assessment scores of schizophrenic in-patients. Neurosci Lett 440: 150–154. Petryshen TL, Middleton FA, Kirby A, Aldinger KA, Purcell S, Tahl AR, Morley CP, McGann L, Gentile KL, Rockwell GN, et al. 2005. Support for involvement of neuregulin 1 in schizophrenia pathophysiology. Mol Psychiatry 10:366–374, 328. Ross DE, Kirkpatrick B, Karkowski LM, Straub RE, MacLean CJ, O’Neill FA, Compton AD, Murphy B, Walsh D, Kendler KS. 2000. Sibling correlation of deficit syndrome in the Irish study of high-density schizophrenia families. Am J Psychiatry 157:1071–1076. Saggers-Gray L, Heriani H, Handoko HY, Irmansyah I, Kusumawardhani AA, Widyawati I, Amir N, Nasrun MW, Schwab SG, Wildenauer DB. 2008. Association of PIP5K2A with schizophrenia: A study in an Indonesian family sample. Am J Med Genet Part B 147B:1310–1313. Sanders AR, Duan J, Levinson DF, Shi J, He D, Hou C, Burrell GJ, Rice JP, Nertney DA, Olincy A, et al. 2008. No significant association of 14 candidate genes with schizophrenia in a large European ancestry sample: Implications for psychiatric genetics. Am J Psychiatry 165: 497–506. Schwab SG, Knapp M, Mondabon S, Hallmayer J, Borrmann-Hassenbach M, Albus M, Lerer B, Rietschel M, Trixler M, Maier W, et al. 2003. Support for association of schizophrenia with genetic variation in the 6p22.3 gene, dysbindin, in sib-pair families with linkage and in an additional sample of triad families. Am J Hum Genet 72:185– 190. Schwab SG, Knapp M, Sklar P, Eckstein GN, Sewekow C, BorrmannHassenbach M, Albus M, Becker T, Hallmayer JF, Lerer B, et al. 2006. Evidence for association of DNA sequence variants in the phosphatidylinositol-4-phosphate 5-kinase IIalpha gene (PIP5K2A) with schizophrenia. Mol Psychiatry 11:837–846. RETHELYI ET AL. 801 Selvin S. 2004. Statistical analysis of epidemiologic data. New York: Oxford University Press. for schizophrenia in the CATIE study: Results of stage 1. Mol Psychiatry 13:570–584. Stefanis NC, Trikalinos TA, Avramopoulos D, Smyrnis N, Evdokimidis I, Ntzani EE, Ioannidis JP, Stefanis CN. 2007. Impact of schizophrenia candidate genes on schizotypy and cognitive endophenotypes at the population level. Biol Psychiatry 62:784–792. Szekeres G, Keri S, Juhasz A, Rimanoczy A, Szendi I, Czimmer C, Janka Z. 2004. Role of dopamine D3 receptor (DRD3) and dopamine transporter (DAT) polymorphism in cognitive dysfunctions and therapeutic response to atypical antipsychotics in patients with schizophrenia. Am J Med Genet Part B 124B:1–5. Stefansson H, Sigurdsson E, Steinthorsdottir V, Bjornsdottir S, Sigmundsson T, Ghosh S, Brynjolfsson J, Gunnarsdottir S, Ivarsson O, Chou TT, et al. 2002. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet 71:877–892. Stopkova P, Saito T, Fann CS, Papolos DF, Vevera J, Paclt I, Zukov I, Stryjer R, Strous RD, Lachman HM. 2003. Polymorphism screening of PIP5K2A: A candidate gene for chromosome 10p-linked psychiatric disorders. Am J Med Genet Part B 123B:50–58. Stopkova P, Vevera J, Paclt I, Zukov I, Papolos DF, Saito T, Lachman HM. 2005. Screening of PIP5K2A promoter region for mutations in bipolar disorder and schizophrenia. Psychiatr Genet 15:223–227. Straub RE, Weinberger DR. 2006. Schizophrenia genes—Famine to feast. Biol Psychiatry 60:81–83. Straub RE, Jiang Y, MacLean CJ, Ma Y, Webb BT, Myakishev MV, HarrisKerr C, Wormley B, Sadek H, Kadambi B, et al. 2002. Genetic variation in the 6p22.3 gene DTNBP1, the human ortholog of the mouse dysbindin gene, is associated with schizophrenia. Am J Hum Genet 71:337– 348. Strauss GP, Harrow M, Grossman LS, Rosen C. 2008. Periods of recovery in deficit syndrome schizophrenia: A 20-year multi-follow-up longitudinal study. Schizophr Bull. DOI: 10.1093/schbul/sbn167. Sullivan PF. 2007. Spurious genetic associations. Biol Psychiatry 61: 1121–1126. Sullivan PF, Kendler KS, Neale MC. 2003. Schizophrenia as a complex trait: Evidence from a meta-analysis of twin studies. Arch Gen Psychiatry 60:1187–1192. Sullivan PF, Lin D, Tzeng JY, van den Oord E, Perkins D, Stroup TS, Wagner M, Lee S, Wright FA, Zou F, et al. 2008. Genomewide association Talkowski ME, Seltman H, Bassett AS, Brzustowicz LM, Chen X, Chowdari KV, Collier DA, Cordeiro Q, Corvin AP, Deshpande SN, et al. 2006. Evaluation of a susceptibility gene for schizophrenia: Genotype based meta-analysis of RGS4 polymorphisms from thirteen independent samples. Biol Psychiatry 60:152–162. Tek C, Kirkpatrick B, Buchanan RW. 2001. A five-year followup study of deficit and nondeficit schizophrenia. Schizophr Res 49:253–260. The International HapMap Consortium. 2005. A haplotype map of the human genome. Nature 437:1299–1320. Vrijenhoek T, Buizer-Voskamp JE, van der Stelt I, Strengman E, Sabatti C, Geurts van Kessel A, Brunner HG, Ophoff RA, Veltman JA. 2008. Recurrent CNVs disrupt three candidate genes in schizophrenia patients. Am J Hum Genet 83:504–510. Williams NM, O’Donovan MC, Owen MJ. 2005. Is the dysbindin gene (DTNBP1) a susceptibility gene for schizophrenia? Schizophr Bull 31:800–805. Wonodi I, Mitchell BD, Stine OC, Hong LE, Elliott A, Kirkpatrick B, Carpenter WT Jr, Thaker GK, Buchanan RW. 2006. Lack of association between COMT gene and deficit/nondeficit schizophrenia. Behav Brain Funct 2:42. Yue W, Kang G, Zhang Y, Qu M, Tang F, Han Y, Ruan Y, Lu T, Zhang J, Zhang D. 2007. Association of DAOA polymorphisms with schizophrenia and clinical symptoms or therapeutic effects. Neurosci Lett 416: 96–100. Zhang F, St Clair D, Liu X, Sun X, Sham PC, Crombie C, Ma X, Wang Q, Meng H, Deng W, et al. 2005. Association analysis of the RGS4 gene in Han Chinese and Scottish populations with schizophrenia. Genes Brain Behav 4:444–448.