Bipolar disorder in the Bulgarian Gypsies Genetic heterogeneity in a young founder population.код для вставкиСкачать
RESEARCH ARTICLE Neuropsychiatric Genetics Bipolar Disorder in the Bulgarian Gypsies: Genetic Heterogeneity in a Young Founder Population Radka Kaneva,1 Vihra Milanova,2 Dora Angelicheva,3 Stuart MacGregor,4 Christian Kostov,2 Rositza Vladimirova,2 Spiridon Aleksiev,2 Mina Angelova,1 Vessela Stoyanova,2 Angeline Loh,3 Joachim Hallmayer,5 Luba Kalaydjieva,3 and Assen Jablensky6* 1 Molecular Medicine Centre and National Genetic Laboratory, Medical University of Sofia, Sofia, Bulgaria Department of Psychiatry, Medical University of Sofia and First Psychiatric Clinic, Alexandrovska University Hospital, Sofia, Bulgaria 2 3 Centre for Medical Research and Western Australian Institute for Medical Research, The University of Western Australia, Perth, Australia 4 Genetic Epidemiology, Queensland Institute of Medical Research, Royal Brisbane Hospital, Brisbane, Australia Department of Psychiatry and Behavior Sciences, Stanford, California 5 6 School of Psychiatry and Clinical Neurosciences, The University of Western Australia, Perth, Australia Received 9 December 2007; Accepted 27 March 2008 We report the results of follow-up analyses of 12 genomic regions showing evidence of linkage in a genome-wide scan (GWS) of Gypsy families with bipolar affective disorder (BPAD). The Gypsies are a young founder population comprising multiple genetically differentiated sub-isolates with strong founder effect and limited genetic diversity. The BPAD families belong to a single sub-isolate and are connected by numerous inter-marriages, resulting in a super-pedigree with 181 members. We aimed to re-assess the positive GWS findings and search for evidence of a founder susceptibility allele after the addition of newly recruited subjects, some changes in diagnostic assignment, and the use of denser genetic maps. Linkage analysis was conducted with SimWalk2, accommodating the full complexity of pedigree structure and using a conservative narrow phenotype definition (BPAD only). Six regions were rejected, while 1p36, 13q31, 17p11, 17q21, 6q24, and 4q31 produced nominally significant results in both the individual families and the super-pedigree. Haplotypes were reconstructed and joint tests for linkage and association were done for the most promising regions. No common ancestral haplotype was identified by sequencing a strong positional and functional candidate gene (GRM1) and additional STR genotyping in the top GWS region, 6q24. The best supported region was a 12 cM interval on 4q31, also implicated in previous studies, where we obtained significant results in the superpedigree using both SimWalk2 (P ¼ 0.004) and joint Pseudomarker analysis of linkage and linkage disequilibrium (P ¼ 0.000056). The size of the region and the characteristics of the Gypsy population make it suitable for LD mapping. Ó 2008 Wiley-Liss, Inc. Key words: bipolar disorder; linkage; genetic isolate; haplotypes; linkage disequilibrium Ó 2008 Wiley-Liss, Inc. How to Cite this Article: Kaneva R, Milanova V, Angelicheva D, MacGregor S, Kostov C, Vladimirova R, Aleksiev S, Angelova M, Stoyanova V, Loh A, Hallmayer J, Kalaydjieva L, Jablensky A. 2009. Bipolar Disorder in the Bulgarian Gypsies: Genetic Heterogeneity in a Young Founder Population. Am J Med Genet Part B 150B:191–201. INTRODUCTION Bipolar affective disorder, BPAD, (lifetime risk 1%; heritability >80%) is a severe mood disorder, characterized by early onset, This article contains supplementary material, which may be viewed at the American Journal of Medical Genetics website at http://www.interscience. wiley.com/jpages/1552-4841/suppmat/index.html. Radka Kaneva, Vihra Milanova, Luba Kalaydjieva, and Assen Jablensky contributed equally to this work. Grant sponsor: National Health and Medical Research Council of Australia; Grant number: 353648; Grant sponsor: National Science Fund, Ministry of Education and Science, Bulgaria; Grant number: G2/2003. Christian Kostov’s present address is Pfalzklinikum, Klinik fuer Psychiatrie und Psychotherapie, Rockenhausen, Germany. *Correspondence to: Prof. Assen Jablensky, School of Psychiatry & Clinical Neurosciences, University of Western Australia, MRF Building, 50 Murray Street, Mail Bag Delivery Point M571, Perth, WA 6000, Australia. E-mail: firstname.lastname@example.org Published online 28 April 2008 in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/ajmg.b.30775 191 192 alternation of manic/hypomanic and depressive episodes, partial or complete remissions between the episodes, significant functional impairment and high risk of suicide [Thomas, 2004; Cheng et al., 2006]. Despite recent attempts at identifying cognitive, neurophysiological or neuroimaging endophenotypes [Bearden and Freimer, 2006; Savitz and Ramesar, 2006; Antila et al., 2007] there is no validated biological marker and the diagnosis rests on clinical observation and detailed history. The genetic basis of BPAD is poorly understood despite general agreement that genetic factors play a major role in susceptibility [Tsuang and Faraone, 2000; Smoller and Finn, 2003]. Individual linkage studies have identified a number of candidate loci [Baron, 2002; Craddock and Forty, 2006], however replication attempts and meta-analyses [Badner and Gershon, 2002; Segurado et al., 2003; McQueen et al., 2005] have produced inconsistent results. While recent genome-wide association studies, made possible by new genotyping technologies and large case–control samples, have been remarkably successful in identifying susceptibility genes for disorders with well defined phenotypes [Easton et al., 2007; Frayling, 2007; Todd et al., 2007; WTCCC, 2007], they have been less productive in BPAD [WTCCC, 2007; Baum et al., 2008], reinforcing the concept of extensive genetic heterogeneity and the need for biologically anchored phenotype definitions. The presumed advantages of founder populations for genetic research include a relatively uniform genetic background, shared environment, phenotype definitions that are easier to standardize, and large intact traditional families. BPAD is among the disorders most investigated in such populations, for example, in the Old Order Amish, Costa Rican, French Canadian, and Finnish, etc. [Pekkarinen et al., 1995; Freimer et al., 1996; Ginns et al., 1996; Morissette et al., 1999; Ekholm et al., 2003; Venken et al., 2005; Herzberg et al., 2006], with a common pattern of failure of follow-up studies to replicate initial evidence of linkage, interand intra-family heterogeneity and lack of shared founder haplotypes [Egeland et al., 1987; Kelsoe et al., 1989; McInnes et al., 1996; Ekholm et al., 2002, 2003; Ophoff et al., 2002; Shink et al., 2003]. In the last decade, genetic studies of the Roma/Gypsies have placed them among the most interesting isolates of Europe [Kalaydjieva et al., 2001, 2005]. The Gypsies are a young founder population of Asian ancestry, whose history is a string of population fissions generating multiple endogamous, culturally divergent groups with genetic profiles shaped by demographic bottlenecks, random drift and differential admixture [Kalaydjieva et al., 2005]. A common primary founder effect, limited diversity, and a substructure resulting from the genetic differentiation of the endogamous Gypsy groups [Gresham et al., 2001; Morar et al., 2004] hold promise for genetic research comparable or even superior to some of the better known isolated populations. In a previous study, we conducted a genome-wide scan and initial fine mapping in three Gypsy families as part of a larger linkage study of BPAD [Schumacher et al., 2005]. In that study, the Gypsy families produced the most promising results, with significant evidence of linkage to 4q31 and at least borderline significance for 1p36-p34 and 6q24 [Schumacher et al., 2005]. Nine further regions showed suggestive evidence of linkage in the Gypsy families. AMERICAN JOURNAL OF MEDICAL GENETICS PART B Here we describe the follow-up investigations, putting our initial positive findings to the standard test of additional family members, changes in affection status and denser genetic maps. In addition, to analyze the data within the intact and complex structure of the three inter-related Gypsy families, we used the MCMC-based (Markov Chain Monte Carlo) linkage program SimWalk2 v.2.91 [Sobel and Lange, 1996]. METHODS The Families The three families (Fig. 1) are part of the Gypsy community, currently numbering 7,000 individuals, in a small town in the southwest of Bulgaria. Similar to the entire Gypsy population of Eastern Europe, this community has been strongly affected by the socio-economic changes of the 1990s, with high unemployment rates, widespread alcoholism and drug abuse, and rising somatic morbidity and mortality. The community belongs to a sub-isolate whose estimated time of founding (based on Y chromosome haplotype diversity) is 300 years or 10–12 generations back, with size of the founding male population estimated at 300 to 500 individuals (Chaix and Kalaydjieva, unpublished work). Our previous population genetic studies of this sub-isolate have shown limited diversity and relative homogeneity [Kalaydjieva et al., 2005]. Since no formal genealogical records are available, we used narrative histories by multiple informants to reconstruct the family trees. Independently confirmed data, going up to 7 generations back, revealed a number of inter-marriages between the three families, leading to a super-pedigree with a total of 181 members (a 3D representation of the super-pedigree can be found at http:// www.waimr.uwa.edu.au/bipolar.html). The small founding size and the traditional marriage patterns suggest that the relatedness between the founders of the affected families is likely to be closer than revealed by the super-pedigree structure. Our study thus focuses a large affected kindred from a young, genetically restricted sub-isolate, within an isolated founder population. Therefore our assumption is of genetic homogeneity and we define the best candidates for further studies in the Gypsy population as regions where evidence of linkage obtained in the analysis of the individual families is supported by the joint pedigree, with allele sharing across families. Ascertainment and Diagnostic Assessment Two of the families were first brought to the attention of the research team by the regional mental health service in 1993. The third affected family was identified in 1996. Family members were interviewed with the Schedules for Clinical Assessment in Neuropsychiatry [Wing et al., 1990] and the Diagnostic Interview for Psychoses [Castle et al., 2006]. Ongoing management of the case finding process over the 14 years of this project, and monitoring of the diagnostic assessment was provided by VM. Case records were provided by local psychiatrists who had followed up the affected individuals from the onset of the disease. Regular contact with the families and with the local mental health services allowed prompt detection of new cases and of changes in KANEVA ET AL. 193 FIG. 1. Pedigrees used in the linkage analysis. the clinical course of the disorder. SCAN/DIP data and case records were independently reviewed by two senior psychiatrists (AJ and VM), blinded to the positions of the subjects in the pedigree. Reference diagnoses, according to DSM-IV and ICD-10 criteria, were generated using the OPCRIT computerized diagnostic algorithm [McGuffin et al., 1991]. Any discrepancies between the OPCRIT diagnoses and the clinicians’ judgment were reviewed until a final research diagnosis (lifetime) was established, using a 194 AMERICAN JOURNAL OF MEDICAL GENETICS PART B TABLE I. Affected Subjects From the Three Gypsy Families Included in the Initial Genome Wide Scan and Fine Mapping [Schumacher et al., 2005] and in the Current Study GS þ initial FM, n (n genotyped) Family 35 41 114 All affected BPAD (BP I þ BP II) 14 (10) 2 (2) 2 (1) 18 (13) This study, n (n genotyped) BPAD þ UPR 26 (21) 8 (8) 7 (6) 41 (35) BPAD (BP I þ BP II) 16 (11) 5 (5) 6 (5) 27 (21) BPAD þ UPR 30 (27) 16 (16) 11 (10) 57 (53) BPAD, bipolar affective disorder; BP I, bipolar disorder, type 1; BPII, Bipolar disorder type 2; UPR, unipolar recurrent depression. The changes in the number of affected individuals are the compound result of newly diagnosed cases, recruitment of additional family branches, and changes in affection status. best estimate consensus procedure based on all available sources of information. A total of 92 living members of the three families participated in the current study, with a larger number of affected subjects compared to the earlier genome-wide linkage scan (Table I). Written informed consent was obtained from all individuals. The project was approved by the human research ethics committees of The University of Western Australia, the Medical University of Sofia and the Alexandrovska University Hospital. Since the previous study [Schumacher et al., 2005], the diagnosis of 11 affected subjects out of total 57 (35 cases included in the genome scan) underwent revision. In 3 cases it was changed from recurrent depressive disorder (RDD) to BPAD, in 2 cases from BPAD to RDD, and in 6 from RDD to unknown. The latter group included 5 individuals affected by epilepsy. Epilepsy, including temporal lobe epilepsy (TLE), was diagnosed in members of some of the family branches and in individual cases scattered across the extended kindred (the super-pedigree at http://www.waimr.uwa. edu.au/bipolar.html). Since depression is common and can be a secondary phenomenon in TLE patients, these 5 subjects were reclassified as unknown. In addition, 2 RDD cases with history of severe head injury, and one with thyroid dysfunction were also reclassified as unknown. In view of such co-morbidity, which could confound the diagnosis of mood disorders, and the concept that for research purposes bipolar and unipolar depressive disorder should be treated as separate entities [Kendell, 1987; Farmer and McGuffin, 1989], this study focuses on bipolar affective disorder. Considering the widespread impact of psychosocial and economic adversity on this community, the numerous cases of depressive disorder in the families were conservatively treated as unknown in the linkage analyses. The total number of family members with BPAD was 27 (21 with BP I and 6 with BP II), of whom 21 were genotyped. The age of the 21 cases of BP I (14 females) at the time of diagnostic assessment was in the range between 17 and 69 years (mean 42.7 years), and the estimated mean age at onset was 28.8 years (<20 years of age in 9 cases and >40 years of age in 5 cases). In the 6 cases of BP II (3 females), the mean age at onset was 23.7 years (<20 years of age in 3 cases and >40 years of age in 1 case). Only 4 patients in the BP I group had ever experienced psychotic symptoms, and the overall course of the disorder was one of moderate severity (ICD-10 criteria). There were no atypical features in either the manic/hypomanic or depressive episodes, and the commonest comorbidity was with anxiety disorder, especially panic attacks. Substance abuse was relatively uncommon, with only 2 male patients diagnosed with alcohol dependence. Suicide attempts (including one consumed suicide) had been recorded in 13 out of the total of 27 patients in the three families. The majority of the patients had experienced multiple affective episodes, with good response to standard carbamazepine treatment and nearly complete remissions between the episodes. Genetic Analyses Genetic markers and genotyping. The analyses included 12 regions on 10 chromosomes, where significant or suggestive NPL values had been obtained in our previous genome-wide scan [Schumacher et al., 2005]. A summary of the regions and the number of markers is given in Supplementary Table I. Microsatellite markers, previously used in the genome scan or first round of fine mapping, were genotyped in the additionally recruited family members, and newly added microsatellites were genotyped in the whole sample. The microsatellites were chosen from the UCSC database (http://genome.ucsc.edu/). PCR amplification was done under standard conditions. Electrophoretic length separation was performed on an ABI377 DNA Analyzer or a Hitachi FMBI0II fluorescent scanner. Genotypes were checked for Mendelian inconsistencies using the PedCheck program [O’Connell and Weeks, 1998]. Inconsistencies were resolved by retyping the markers for the specific subjects. Sequencing analysis of GRM1. The coding sequence of GRM1 gene (Ref Seq NM_000838) including the 50 and 30 UTRs, all exons and the exon–intron boundaries, was covered by 23 PCR fragments ranging from 270 to 543 bp (primer sequences available upon request). Sequencing of the PCR products was performed following the standard protocols for BigDye 3.1 and was analyzed on ABI 377 DNA Sequencer. Linkage analysis. We used the integrated genetic map (available at www.qimr.edu.au/davidd/davidd.html). Allele frequencies were calculated by maximum likelihood estimates (MLE) using the full pedigree structure. These estimates impute the marker genotypes for untyped individuals and take into account the pedigree structure. The estimates were computed in SOLAR, which in turn uses the external program Allfreq, an extension of the MENDEL software [Almasy and Blangero, 1998]. Input files were generated with the Mega2 Version 3.0 software [Mukhopadhyay KANEVA ET AL. 195 et al., 2005]. Non-parametric linkage analysis was performed using SimWalk2 v.2.91 [Sobel and Lange, 1996; Sobel et al., 2001]. SimWalk uses a Markov Chain Monte Carlo (MCMC) approach to sample from the complete distribution of underlying inheritance patterns proportional to their likelihood, which is calculated from the observed genotype data. Empirical P values are obtained by comparing the observed value of the statistics to that estimated under the null hypothesis by repeated sampling of marker data simulated with a gene dropping algorithm without linkage to phenotype. Haplotypes were reconstructed using the HAPLOTYPE module of SimWalk2, under the assumption of linkage equilibrium between markers. To test jointly for linkage and association, we used PSEUDOMARKER, a linkage analysis software which is more powerful than the conventional ‘‘model-free’’ tests when both linkage and LD are present and remains powerful for detection of either of the two phenomena in the absence of the other [Goring and Terwilliger, 2000]. RESULTS Out of the 12 regions selected for follow-up analysis, 6 met our criterion of positive results in both the individual families and the combined super-pedigree with nominally significant results (P < 0.05, log 10(p) > 1.3). The data are summarized in Table II. Regions Not Supported by the Follow-Up Analyses The 8p21-p12, 9p24-p22, 11p13-p11, and 18q12-q21 regions did not show significant results in any of the SimWalk2 statistics, neither in the individual families nor in the combined super-family. In two chromosome 2 regions, 2p23-p13 (D2S352–D2S286) and 2q21-q31 (D2S368–D2S376) we re-analyzed the previous genome scan data [Schumacher et al., 2005] using SimWalk2 and the full family structure. Positive results were obtained only for the 14 cM interval on 2q (D2S368–D2S2275) when the three families were analyzed separately. The nominally significant values in the NPL_ALL statistic, with a maximum of log 10(p) ¼ 1.88 at D2S2367, were contributed by one of the three families. The analysis of the combined super-family failed to produce any positive results. Weakly Positive Regions Four regions with initially suggestive evidence of linkage remained nominally significant in the follow-up analysis: 1p36-p34, 13q22q32, 17p11-q12, and 17q21-q24. A total of 10 markers were analyzed on 1p36-1p34, three newly added to the D1S2732–D1S493 interval and seven from the previous fine mapping study [Schumacher et al., 2005]. Nominally significant results were obtained in the analysis of both the individual families and the combined pedigree over the 6 cM D1S458–D1S2693 interval. The highest values were produced by the NPL_PAIR statistic in the analysis of the joined superpedigree at markers D1S2885 (log 10(p) ¼ 1.65) and D1S2749 (log 10(p) ¼ 1.61). The fine mapping of chromosome 13 was done with nine STR markers. Nominally significant P values were obtained for several statistics and markers in the region between D13S1804 and D13S265, encompassing 9 cM. The highest results were observed at D13S170 (log 10(p) ¼ 1.93) for NPL_PAIR in the families analyzed separately. A lower but still significant value was obtained at the same marker and the neighboring D13S790, both showing log 10(p) ¼ 1.4 for NPL_PAIR in the joint super-pedigree. Two regions were analyzed on chromosome 17, where a total of 11 markers were genotyped. On 17p11-q12, the best values were obtained in the analysis of the joint super-pedigree with the BLOCKS statistic (based on the number of different founder alleles contributing to the affected) over a 5 cM region, with a maximum TABLE II. Results of the SimWalk2 Linkage Analyses in the Three Separate Gypsy Families and in the Joint Super-Pedigree This study Chr 1 2 4 6 8 9 11 13 17 17 18 Genome scan and initial fine mapping resultsa þ þ þþ þþ þ þ þ þ þ þ þ Three ‘‘unrelated’’ families Significance (P)b 0.03 0.01 0.004 0.003 NS NS NS 0.012 0.03 0.03 NS Marker region D1S478–D1S2639 D2S368–D2S2275 D4S420–D4S2999 GATA184A08–D6S494 Interval (cM) 39.22–47.61 147.49–161.26 137.57–149.32 150.28–155.85 D13S1804–D13S265 D17S2196–D17S1294 D17S1820–D17S790 72.66–80.80 47.40–52.78 78.15–82.15 One joint super-pedigree Significance (P)b 0.01 NS 0.007 0.02 NS NS NS 0.04 0.03 0.04 NS Marker region D1S458-D1S2639 Interval (cM) 41.81–47.61 D4S420–D4S1606 GATA184A08–D6S494 137.57–145.12 150.28–155.75 D13S1812–D13S265 D17S2196–D17S1294 D17S1820–D17S790 74.15–80.80 47.40–52.78 78.15–82.15 NS, non-significant (P > 0.05). a Suggestive genome scan and 1st round fine mapping data are denoted with ‘‘þ’’, while results at significant or close to significant level with ‘‘þþ’’. b Scores obtained with SimWalk2 [Sobel and Lange, 1996]. The bolded figures are the significance values. 196 AMERICAN JOURNAL OF MEDICAL GENETICS PART B FIG. 2. SimWalk2 analysis (MAX-TREE and NPL_PAIR statistics) of chromosome 6q24 in the three separate Gypsy families and in the joint superpedigree. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.] at D17S1871 (log 10(p) ¼ 1.69). On 17q21-q24, the best supported area spans 4 cM between D17S1820 and D17S790, with the highest result at marker D17S1820, again for the joint superpedigree and the BLOCKS statistic (log 10(p) ¼ 1.5). The 6q23-q24 region was a strong candidate for follow-up. The NPL values at 6q24 were the highest in the genome-wide scan and initial fine mapping, with maximum scores of 3.46 at 147.75 cM and 4.87 at 152.00 cM, suggesting the possibility of more than one susceptibility gene on 6q [Schumacher et al., 2005] The follow-up SimWalk2 analysis was performed on a set of nine markers (Fig. 2), with an average inter-marker distance of 3.08 cM. The analysis of the three separate families supported the telomeric peak, between markers GATA184A08 and D6S494 (Fig. 2) with both the MAXTREE and NPL_PAIR statistic producing P values < 0.01. The maximum (log 10(p) ¼ 2.53) was obtained at D6S960 for the NPL pair statistic. All families contributed to these results. In the combined super-family, the findings were less convincing with a maximum (log 10(p) ¼ 1.58) produced by NPL_PAIR still at D6S960. Haplotype analysis showed three segments that were shared by different subsets of affected individuals from the three families: a 3-marker haplotype D6S1587–D6S403–D6S970 haplotype (alleles 3-4-4) and two 2-marker haplotypes formed by D6S970–GATA184A08 (alleles 5-2) and by D6S960 and D6S494 (alleles 5-1). These regions were examined further, with the addition of microsatellite markers and sequencing of a strong candidate gene, the Glutamate Receptor Metabotropic 1 (GRM1). Five additional microsatellites were genotyped in the D6S1587–D6S970 interval, seven in the D6S970–GATA184A08 interval (also containing GRM1), and one between D6S960 and D6S494. In all three regions, the additional markers revealed that the seemingly conserved 2- and 3-marker haplotypes were in reality non-identical, with diversification introduced by the newly added microsatellites, as well as by the SNPs in GRM1. The addition of five microsatellites to the 4.8 cM interval D6S1587–D6S403–D6S970 specified four unrelated haplotypes. The sequencing analysis of the coding regions, flanking introns and UTRs of GRM1 identified polymorphic nucleotide positions only in the largest exon 8. All were common polymorphisms represented in dbSNP (www.ncbi.nlm.nih.gov/ projects/snp/). Their analysis in the chromosomes that carried the same (5-2) D6S970–GATA184A08 haplotype, revealed four different haplotypes over the 0.9 kb sequence of GRM1 exon 8 (Fig. 3). In the weakly supported telomeric region (152 cM), the addition of marker D6S1687 in-between D6S960 and D6S494 led to the differentiation of the 5-1 haplotype into three haplotypes: 5-4-1, 5-3-1, and 5-5-1. Best Supported Region Chromosome 4q28-q32 (D4S1615–D4S2982) was the most promising result emerging from the genome scan and the following initial fine mapping in the Gypsy families [Schumacher et al., 2005]. In the present study, the large linked region (28 cM) was saturated with a total of 24 markers genotyped in the enlarged family sample. Both the analyses in the three separate families and in the joint superpedigree produced significant results for the BLOCKS and NPL_PAIR statistics. The best supported linkage area was between markers D4S420 and D4S2976 (Fig. 4), reducing the size of the region of interest to 12 cM. The highest value obtained in the analysis of individual families was at marker D4S2999 KANEVA ET AL. 197 FIG. 3. Haplotype analysis of GRM1 Exon 8 on Chromosome 6q24, located within the ‘‘conserved’’ haplotype D6S970–GATA184A08 (5-2). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.] (log 10(p) ¼ 2.4, P ¼ 0.004) for the NPL_PAIR. The joint superpedigree showed significant results for all markers in the D4S420–D4S2962 region (log 10(p) 2, P < 0.01) for the BLOCKS statistic. PSEUDOMARKER also produced positive two-point LOD scores with a maximum of 3.18 at D4S2999 under a recessive model (Table III). The PSEUDOMARKER [Goring and Terwilliger, 2000] analysis showed in addition evidence of association, with P-values for some of the markers ranging between 0.01 and 0.03 (Table III), which however did not withstand correction for multiple testing. Haplotype analysis identified the 4-4 haplotype, formed by markers D4S2999 and D4D2976, as significantly different in frequency between affected and unaffected individuals in all three families (44.44% of chromosomes in affected subjects and 26.74% in unaffected; Fisher’s exact test P ¼ 0.019). The D4S3014–D4S2962 (4-2) haplotype, by contrast, was more common in unaffected family members (22.5%), compared to just 7.6% in affected (Fisher’s exact test P ¼ 0.019). DISCUSSION The main aim of the current study was to re-assess the BPAD linkage findings in three large inter-related Gypsy families. Since the genome scan producing these findings [Schumacher et al., 2005], our sample has undergone the kind of changes that have contributed to non-replication in other studies, such as addition of more family members and re-evaluation of affection status. We FIG. 4. SimWalk2 analysis (BLOCKS and NPL_PAIR statistics) of chromosome 4q31 in the three separate Gypsy families and in the joint superpedigree. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.] 198 AMERICAN JOURNAL OF MEDICAL GENETICS PART B TABLE III. PSEUDOMARKER Analysis for Linkage and Association on 4q31 Under a Recessive Model Marker D4S420 D4S1644 D4S424 D4S1625 D4S2981 D4S1604 D4S1586 D4S3014 D4S2962 D4S1606 D4S1548 D4S1588 D4S2934 D4S2999 D4S2976 Two-point LOD 1.63 0.70 0.46 2.78 1.31 0.26 0.50 1.04 2.59 0.38 0.50 2.24 1.62 3.18 0.30 Linkage 0.003048 0.035829 0.073771 0.000172 0.006951 0.134938 0.066555 0.014312 0.000276 0.091799 0.063856 0.000665 0.003201 0.000065 0.118645 LD|Linkage 0.249896 0.035973 0.172117 0.043070 0.531488 0.166867 0.016871 0.381349 0.779549 0.427078 0.031700 0.083871 0.041856 0.011476 0.086875 Linkage|LD 0.017944 0.002520 0.059955 0.000007 0.007387 0.016464 0.008227 0.022948 0.000400 0.516976 0.026122 0.000183 0.002933 0.000004 0.092412 LD þ linkage 0.027700 0.014375 0.111223 0.000391 0.061843 0.139180 0.010016 0.085804 0.019196 0.298751 0.018875 0.002333 0.002762 0.000056 0.070057 The bolded figures are the significance values. have also adopted a conservative approach to linkage analysis by assigning all unaffected individuals to the ‘‘unknown’’ category and by using a statistical package which takes into account, and can handle, the true and complete family structure. Based on the characteristics of the sample and of the population from which it is derived, we searched for evidence of sharing of ancestral alleles between affected subjects to define candidate regions likely to be productive in future studies of BPAD in the same and genetically related sub-isolates. The rejection of the 8p21-p12, 9p24-p22, 11p13-p11, and 18q12q21, based on negative results in both individual families and the super-pedigree, could be due to the changes introduced at this stage of the study, or simply to false positive previous findings. The new data on chromosome 2, produced by re-analysis (without additional genotyping) of the initial data on 2p23-p13 and 2q21-q31, also failed to meet the homogeneity and sharing criteria. Weak evidence was obtained for the 2q locus in the separate families, but no significant values were produced by the joint pedigree. While a recent re-analysis of the family sample used in our initial linkage study [Schumacher et al., 2005] suggests that epistatic interactions between the 2q22-q24 and 6q23-q24 loci play a role in susceptibility to BPAD [Abou Jamra et al., 2007], our results do not support 2q (as well as 6q, see below) as candidate regions for further studies in the Gypsy population. Chromosomes 1p36-p34, 13q22-q32, 17p11-q12, and 17q21q24 remained weakly positive, with nominally significant P values (0.05–0.01) obtained in both the individual families and in the joint super-pedigree. These regions have been implicated in previous studies [Detera-Wadleigh et al., 1999; Cichon et al., 2001; Ewald et al., 2003; Liu et al., 2003; Potash et al., 2003; McGuffin et al., 2005] in other populations and cannot be rejected in the Gypsy population. Their further investigation will require much larger sample sizes. The results of our 6q analysis tell a cautionary tale. The long arm of chromosome 6 has attracted a lot of interest in psychiatric genetics, with positive findings in linkage and association studies of both affective disorder and schizophrenia [Kohn and Lerer, 2005] and is one of the few regions supported by meta-analysis [McQueen et al., 2005]. At least five sub-regions have been implicated in the different studies, making it difficult to assess the significance of these findings. In the genome scan of the Gypsy families, the best NPL result of 3.65 was obtained for a 13 cM region on 6q24, where the initial fine mapping [Schumacher et al., 2005] outlined two peaks, at 147.75 and 152 cM. Analyzed separately and broken down into smaller pedigrees that can be handled by standard linkage packages, all three Gypsy families contributed to these results, with no evidence of heterogeneity. Overlapping positive findings were reported from a study of BPAD families in an isolated Swedish population, with a LOD-1 region between D6S310 and D6S1654 (144–147 cM) [Venken et al., 2005], and a genome scan of Arab schizophrenia families, with the linkage region overlapping the centromeric peak in our study [Lerer et al., 2003]. Our follow-up analysis of the separate families supported the previously identified peak(s), however the combined pedigree provided only weak support limited to the telomeric region (152 cM) (Fig. 2). We did not identify an extended ancestral haplotype and the fragmented 2- or 3-marker haplotypes, in different parts of the region and shared by different combinations of affected members of the three families, did not support the concept of a founder mutation. We allowed for the possibility of an old variant, where haplotype decay via recombinations and microsatellite mutations could have resulted in a small conserved region. Our previous studies have shown that, even in the case of rare Mendelian disorders, the size of the conserved ancestral segment surrounding a Gypsy founder mutation can be limited to 150 kb [Kalaydjieva et al., 2000; Varon et al., 2003] and studies in the Finnish population have documented a smaller size of the conserved segment around common susceptibility alleles compared to single-gene disorders [Hovatta et al., 1998; Peltonen, 2000]. Such a scenario was ruled out by the genotyping of additional markers, microsatellites, and SNPs, which led to the disintegration of the seemingly shared segments into distinctly different haplotypes. KANEVA ET AL. The initially most promising locus at 4q31 remained so in the follow-up study. In the genome scan and first round of fine mapping, we obtained significant evidence of linkage with an NPL of 5.49 at 148.4 cM. The MOD-score analysis of all families included in that study [Schumacher et al., 2005] showed the highest value of 4.24 at 147.2 cM for a phenotype including BP I and BP II as affected, an estimated disease allele frequency of 0.045 and a near-additive mode of inheritance. Our follow-up examinations provide further support for this region, with significant results obtained in both the three separate families and the joint super-pedigree for the BLOCKS and NPL_PAIR statistics. The initial large (28 cM) size of the linked interval was reduced to the 12 cM between D4S420 and D4S2976 (Fig. 4). The joint super-pedigree showed significant results for all markers in the D4S420–D4S2962 region (log 10(p) above 2, P < 0.01) for the BLOCKS statistic, supported by the PSEUDOMARKER results (Table III) which also favored a recessive model. The best value in the separate families (P ¼ 0.004) was at the more telomeric marker D4S2999. In the telomeric 0.65 cM part of the region, we also observed a 2-marker (D4S2999-D4S2976) haplotype present in a large proportion of the chromosomes of affected subjects (44.44%), compared to 26.74% of unaffected (P ¼ 0.019). The 4q31 region thus appears to be the best candidate locus, meeting the criteria of this study. This region has been implicated in previous studies of different psychiatric disorders. In the genome scan of 41 Finish families with BPAD, the best evidence for linkage was produced at D4S1629 (151.84 cM) [Ekholm et al., 2003]. In a genome scan of 65 multiplex bipolar families under broad diagnostic criteria including unipolar recurrent depression, suggestive evidence of linkage (NPL 2.8) was obtained for a 10 cM region around D4S1629 [McInnis et al., 2003]. In the investigation of 40 American BPAD families of European and Israeli ancestry, a two-point LOD score of 3.16 was obtained at marker D4S1625 (140.08 cM) [Liu et al., 2003]. Evidence of linkage of 4q31 with other mental disorders includes schizophrenia in an isolated Finish population [Hovatta et al., 1999] and anxiety disorder in 19 extended American families [Kaabi et al., 2006]. Our broad 4q31 region of linkage contains 12 positional and functional candidate genes, for example, the lithium-sensitive phosphatidylinositol calcium second messenger system (inositol polyphosphate-4-phosphatase, type II, INPP4B), the endothelin receptor type A, (EDNRA), genes encoding Ca2þ binding and signaling cascades (TBC1D9 and DCHS2), and genes related to ubiquitination (SMAD1 and the ubiquitin-specific protease USP38), shown by gene expression studies to be involved in the pathophysiology of BPAD [Ryan et al., 2006]. The narrow candidate region, where we have observed haplotype sharing, contains only one gene, DCHS2. A systematic screening of single nucleotide polymorphisms in the region is underway. ACKNOWLEDGMENTS Special thanks to Dr. T. Milenska, a senior psychiatrist in the town of Kyustendil, Bulgaria, who provided invaluable information on the pedigrees and the clinical history of the affected members. We thank all the patients and families for participation and George Onchev and Amelia Nikolova-Hill for initial contribution with 199 family recruitment. The assistance of Vesselin Chorbov with DNA isolation, Michael Hunter and Kate Hollingsworth with the genotyping, and Veneta Popova with the preparation of figures is gratefully acknowledged. The study was supported by project grant 353648 of the National Health and Medical Research Council of Australia and research grant G2/2003 from the National Science Fund, Ministry of Education and Science, Bulgaria. REFERENCES Abou Jamra R, Fuerst R, Kaneva R, Orozco Diaz G, Rivas F, Mayoral F, Gay E, Sans S, Gonzalez MJ, Gil S, et al. 2007. The first genomewide interaction and locus-heterogeneity linkage scan in bipolar affective disorder: Strong evidence of epistatic effects between loci on chromosomes 2q and6q. Am J Hum Genet 81(5):974–986. Almasy L, Blangero J. 1998. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet 62(5):1198–1211. Antila M, Tuulio-Henriksson A, Kieseppa T, Eerola M, Partonen T, Lonnqvist J. 2007. Cognitive functioning in patients with familial bipolar I disorder and their unaffected relatives. Psychol Med 37(5): 679–687. Badner JA, Gershon ES. 2002. Meta-analysis of whole-genome linkage scans of bipolar disorder and schizophrenia. Mol Psychiatry 7(4): 405–411. Baron M. 2002. Manic-depression genes and the new millennium: Poised for discovery. Mol Psychiatry 7(4):342–358. Baum AE, Akula N, Cabanero M, Cardona I, Corona W, Klemens B, Schulze TG, Cichon S, Rietschel M, Nothen MM, et al. 2008. A genomewide association study implicates diacylglycerol kinase eta (DGKH) and several other genes in the etiology of bipolar disorder. Mol Psychiatry 13(2):197–207. Bearden CE, Freimer NB. 2006. Endophenotypes for psychiatric disorders: Ready for primetime? Trends Genet 22(6):306–313. Castle DJ, Jablensky A, McGrath JJ, Carr V, Morgan V, Waterreus A, Valuri G, Stain H, McGuffin P, Farmer A. 2006. The diagnostic interview for psychoses (DIP): Development, reliability and applications. Psychol Med 36(1):69–80. Cheng R, Juo SH, Loth JE, Nee J, Iossifov I, Blumenthal R, Sharpe L, Kanyas K, Lerer B, Lilliston B, et al. 2006. Genome-wide linkage scan in a large bipolar disorder sample from the National Institute of Mental Health genetics initiative suggests putative loci for bipolar disorder, psychosis, suicide, and panic disorder. Mol Psychiatry 11(3):252–260. Cichon S, Schumacher J, Muller DJ, Hurter M, Windemuth C, Strauch K, Hemmer S, Schulze TG, Schmidt-Wolf G, Albus M, et al. 2001. A genome screen for genes predisposing to bipolar affective disorder detects a new susceptibility locus on8q. Hum Mol Genet 10(25):2933–2944. Craddock N, Forty L. 2006. Genetics of affective (mood) disorders. Eur J Hum Genet 14(6):660–668. Detera-Wadleigh SD, Badner JA, Berrettini WH, Yoshikawa T, Goldin LR, Turner G, Rollins DY, Moses T, Sanders AR, Karkera JD, et al. 1999. A high-density genome scan detects evidence for a bipolar-disorder susceptibility locus on 13q32 and other potential loci on 1q32 and 18p11.2. Proc Natl Acad Sci USA 96(10):5604–5609. Easton DF, Pooley KA, Dunning AM, Pharoah PD, Thompson D, Ballinger DG, Struewing JP, Morrison J, Field H, Luben R, et al. 2007. Genomewide association study identifies novel breast cancer susceptibility loci. Nature 447(7148):1087–1093. Egeland JA, Gerhard DS, Pauls DL, Sussex JN, Kidd KK, Allen CR, Hostetter AM, Housman DE. 1987. Bipolar affective disorders linked to DNA markers on chromosome 11. Nature 325(6107):783–787. 200 Ekholm JM, Pekkarinen P, Pajukanta P, Kieseppa T, Partonen T, Paunio T, Varilo T, Perola M, Lonnqvist J, Peltonen L. 2002. Bipolar disorder susceptibility region on Xq24-q27.1 in Finnish families. Mol Psychiatry 7(5):453–459. Ekholm JM, Kieseppa T, Hiekkalinna T, Partonen T, Paunio T, Perola M, Ekelund J, Lonnqvist J, Pekkarinen-Ijas P, Peltonen L. 2003. Evidence of susceptibility loci on 4q32 and 16p12 for bipolar disorder. Hum Mol Genet 12(15):1907–1915. Ewald H, Kruse TA, Mors O. 2003. Genome wide scan using homozygosity mapping and linkage analyses of a single pedigree with affective disorder suggests oligogenic inheritance. Am J Med Genet Part B 120B(1):63–71. Farmer A, McGuffin P. 1989. The classification of the depressions. Contemporary confusion revisited. Br J Psychiatry 155:437–443. Frayling TM. 2007. Genome-wide association studies provide new insights into type 2 diabetes aetiology. Nat Rev Genet 8(9):657–662. AMERICAN JOURNAL OF MEDICAL GENETICS PART B affective disorder in the Old Order Amish. Nature 342(6247):238– 243. Kendell RE. 1987. Diagnosis and classification of functional psychoses. Br Med Bull 43(3):499–513. Kohn Y, Lerer B. 2005. Excitement and confusion on chromosome 6q: The challenges of neuropsychiatric genetics in microcosm. Mol Psychiatry 10(12):1062–1073. Lerer B, Segman RH, Hamdan A, Kanyas K, Karni O, Kohn Y, Korner M, Lanktree M, Kaadan M, Turetsky N, et al. 2003. Genome scan of Arab Israeli families maps a schizophrenia susceptibility gene to chromosome 6q23 and supports a locus at chromosome 10q24. Mol Psychiatry 8(5):488–498. Liu J, Juo SH, Dewan A, Grunn A, Tong X, Brito M, Park N, Loth JE, Kanyas K, Lerer B, et al. 2003. Evidence for a putative bipolar disorder locus on 2p13-16and other potential loci on 4q31, 7q34, 8q13, 9q31, 10q21-24, 13q32, 14q21 and 17q11-12. Mol Psychiatry. 8(3):333–342. Freimer NB, Reus VI, Escamilla MA, McInnes LA, Spesny M, Leon P, Service SK, Smith LB, Silva S, Rojas E, et al. 1996. Genetic mapping using haplotype, association and linkage methods suggests a locus for severe bipolar disorder (BPI) at 18q22-q23. Nat Genet 12(4):436–441. McGuffin P, Farmer A, Harvey I. 1991. A polydiagnostic application of operational criteria in studies of psychotic illness. Development and reliability of the OPCRIT system. Arch Gen Psychiatry 48(8):764–770. Ginns EI, Ott J, Egeland JA, Allen CR, Fann CS, Pauls DL, Weissenbachoff J, Carulli JP, Falls KM, Keith TP, et al. 1996. A genome-wide search for chromosomal loci linked to bipolar affective disorder in the Old Order Amish. Nat Genet 12(4):431–435. McGuffin P, Knight J, Breen G, Brewster S, Boyd PR, Craddock N, Gill M, Korszun A, Maier W, Middleton L, et al. 2005. Whole genome linkage scan of recurrent depressive disorder from the depression network study. Hum Mol Genet 14(22):3337–3345. Goring HH, Terwilliger JD. 2000. Linkage analysis in the presence of errors IV: Joint pseudomarker analysis of linkage and/or linkage disequilibrium on a mixture of pedigrees and singletons when the mode of inheritance cannot be accurately specified. Am J Hum Genet 66(4):1310–1327. McInnes LA, Escamilla MA, Service SK, Reus VI, Leon P, Silva S, Rojas E, Spesny M, Baharloo S, Blankenship K, et al. 1996. A complete genome screen for genes predisposing to severe bipolar disorder in two Costa Rican pedigrees. Proc Natl Acad Sci USA 93(23):13060–13065. Gresham D, Morar B, Underhill PA, Passarino G, Lin AA, Wise C, Angelicheva D, Calafell F, Oefner PJ, Shen P, et al. 2001. Origins and divergence of the Roma (gypsies). Am J Hum Genet 69(6):1314–1331. McInnis MG, Lan TH, Willour VL, McMahon FJ, Simpson SG, Addington AM, MacKinnon DF, Potash JB, Mahoney AT, Chellis J, et al. 2003. Genome-wide scan of bipolar disorder in 65 pedigrees: Supportive evidence for linkage at 8q24, 18q22, 4q32, 2p12, and 13q12. Mol Psychiatry 8(3):288–298. Herzberg I, Jasinska A, Garcia J, Jawaheer D, Service S, Kremeyer B, Duque C, Parra MV, Vega J, Ortiz D, et al. 2006. Convergent linkage evidence from two Latin-American population isolates supports the presence of a susceptibility locus for bipolar disorder in 5q31–34. Hum Mol Genet 15(21):3146–3153. Hovatta I, Lichtermann D, Juvonen H, Suvisaari J, Terwilliger JD, Arajarvi R, Kokko-Sahin ML, Ekelund J, Lonnqvist J, Peltonen L. 1998. Linkage analysis of putative schizophrenia gene candidate regions on chromosomes 3p, 5q, 6p, 8p, 20p and 22q in a population-based sampled Finnish family set. Mol Psychiatry 3(5):452–457. Hovatta I, Varilo T, Suvisaari J, Terwilliger JD, Ollikainen V, Arajarvi R, Juvonen H, Kokko-Sahin ML, Vaisanen L, Mannila H, et al. 1999. A genomewide screen for schizophrenia genes in an isolated Finnish subpopulation, suggesting multiple susceptibility loci. Am J Hum Genet 65(4):1114–1124. Kaabi B, Gelernter J, Woods SW, Goddard A, Page GP, Elston RC. 2006. Genome scan for loci predisposing to anxiety disorders using a novel multivariate approach: Strong evidence for a chromosome 4 risk locus. Am J Hum Genet 78(4):543–553. Kalaydjieva L, Gresham D, Gooding R, Heather L, Baas F, de Jonge R, Blechschmidt K, Angelicheva D, Chandler D, Worsley P, et al. 2000. Nmyc downstream-regulated gene 1 is mutated in hereditary motor and sensory neuropathy-Lom. Am J Hum Genet 67(1):47–58. Kalaydjieva L, Gresham D, Calafell F. 2001. Genetic studies of the Roma (Gypsies): A review. BMC Med Genet 2:5. McQueen MB, Devlin B, Faraone SV, Nimgaonkar VL, Sklar P, Smoller JW, Abou Jamra R, Albus M, Bacanu SA, Baron M, et al. 2005. Combined analysis from eleven linkage studies of bipolar disorder provides strong evidence of susceptibility loci on chromosomes 6q and8q. Am J Hum Genet 77(4):582–595. Morar B, Gresham D, Angelicheva D, Tournev I, Gooding R, Guergueltcheva V, Schmidt C, Abicht A, Lochmuller H, Tordai A, et al. 2004. Mutation history of the roma/gypsies. Am J Hum Genet 75(4):596–609. Morissette J, Villeneuve A, Bordeleau L, Rochette D, Laberge C, Gagne B, Laprise C, Bouchard G, Plante M, Gobeil L, et al. 1999. Genome-wide search for linkage of bipolar affective disorders in a very large pedigree derived from a homogeneous population in Quebec points to a locus of major effect on chromosome 12q23-q24. Am J Med Genet 88(5): 567–587. Mukhopadhyay N, Almasy L, Schroeder M, Mulvihill WP, Weeks DE. 2005. Mega2: Data-handling for facilitating genetic linkage and association analyses. Bioinformatics 21(10):2556–2557. O’Connell JR, Weeks DE. 1998. PedCheck: A program for identification of genotype incompatibilities in linkage analysis. Am J Hum Genet 63(1):259–266. Kalaydjieva L, Morar B, Chaix R, Tang H. 2005. A newly discovered founder population: The Roma/Gypsies. Bioessays 27(10):1084–1094. Ophoff RA, Escamilla MA, Service SK, Spesny M, Meshi DB, Poon W, Molina J, Fournier E, Gallegos A, Mathews C, et al. 2002. Genomewide linkage disequilibrium mapping of severe bipolar disorder in a population isolate. Am J Hum Genet 71(3):565–574. Kelsoe JR, Ginns EI, Egeland JA, Gerhard DS, Goldstein AM, Bale SJ, Pauls DL, Long RT, Kidd KK, Conte G, et al. 1989. Re-evaluation of the linkage relationship between chromosome 11p loci and the gene for bipolar Pekkarinen P, Terwilliger J, Bredbacka PE, Lonnqvist J, Peltonen L. 1995. Evidence of a predisposing locus to bipolar disorder on Xq24-q27.1 in an extended Finnish pedigree. Genome Res 5(2):105–115. KANEVA ET AL. Peltonen L. 2000. Positional cloning of disease genes: Advantages of genetic isolates. Hum Hered 50(1):66–75. Potash JB, Zandi PP, Willour VL, Lan TH, Huo Y, Avramopoulos D, Shugart YY, MacKinnon DF, Simpson SG, McMahon FJ, et al. 2003. Suggestive linkage to chromosomal regions 13q31 and 22q12 in families with psychotic bipolar disorder. Am J Psychiatry 160(4):680–686. Ryan MM, Lockstone HE, Huffaker SJ, Wayland MT, Webster MJ, Bahn S. 2006. Gene expression analysis of bipolar disorder reveals downregulation of the ubiquitin cycle and alterations in synaptic genes. Mol Psychiatry 11(10):965–978. Savitz JB, Ramesar RS. 2006. Personality: Is it a viable endophenotype for genetic studies of bipolar affective disorder? Bipolar Disord 8(4): 322–337. Schumacher J, Kaneva R, Jamra RA, Diaz GO, Ohlraun S, Milanova V, Lee YA, Rivas F, Mayoral F, Fuerst R, et al. 2005. Genomewide scan and finemapping linkage studies in four European samples with bipolar affective disorder suggest a new susceptibility locus on chromosome 1p35-p36 and provides further evidence of loci on chromosome 4q31 and 6q24. Am J Hum Genet 77(6):1102–1111. 201 Sobel E, Lange K. 1996. Descent graphs in pedigree analysis: Applications to haplotyping, location scores, and marker-sharing statistics. Am J Hum Genet 58(6):1323–1337. Sobel E, Sengul H, Weeks DE. 2001. Multipoint estimation of identity-bydescent probabilities at arbitrary positions among marker loci on general pedigrees. Hum Hered 52(3):121–131. Thomas P. 2004. The many forms of bipolar disorder: A modern look at an old illness. J Affect Disord 79(Suppl 1): S3–8. Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K, Plagnol V, Bailey R, Nejentsev S, Field SF, Payne F, et al. 2007. Robust associations of four new chromosome regions from genome-wide analyses of type 1 diabetes. Nat Genet 39(7):857–864. Tsuang MT, Faraone SV. 2000. The future of psychiatric genetics. Curr Psychiatry Rep 2(2):133–136. Varon R, Gooding R, Steglich C, Marns L, Tang H, Angelicheva D, Yong KK, Ambrugger P, Reinhold A, Morar B, et al. 2003. Partial deficiency of the C-terminal-domain phosphatase of RNA polymerase II is associated with congenital cataracts facial dysmorphism neuropathy syndrome. Nat Genet 35(2):185–189. Segurado R, Detera-Wadleigh SD, Levinson DF, Lewis CM, Gill M, Nurnberger JI Jr, Craddock N, DePaulo JR, Baron M, Gershon ES, et al. 2003. Genome scan meta-analysis of schizophrenia and bipolar disorder. Part III. Bipolar disorder. Am J Hum Genet 73(1):49–62. Venken T, Claes S, Sluijs S, Paterson AD, van Duijn C, Adolfsson R, Del-Favero J, Van Broeckhoven C. 2005. Genomewide scan for affective disorder susceptibility Loci in families of a northern Swedish isolated population. Am J Hum Genet 76(2):237–248. Shink E, Morissette J, Barden N. 2003. Genetic heterogeneity in a very large bipolar affective disorder pedigree from Quebec. Am J Med Genet Part B 119B(1):65–68. Wing JK, Babor T, Brugha T, Burke J, Cooper JE, Giel R, Jablenski A, Regier D, Sartorius N. 1990. SCAN. Schedules for Clinical Assessment in Neuropsychiatry. Arch Gen Psychiatry 47(6):589–593. Smoller JW, Finn CT. 2003. Family, twin, and adoption studies of bipolar disorder. Am J Med Genet Part C 123C(1):48–58. WTCCC. 2007. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 447(7145):661–678.