Analysis of microsatellite markers and single nucleotide polymorphisms in candidate genes for susceptibility to bipolar affective disorder in the chromosome 12Q24.31 regionкод для вставкиСкачать
American Journal of Medical Genetics Part B (Neuropsychiatric Genetics) 135B:50 – 58 (2005) Analysis of Microsatellite Markers and Single Nucleotide Polymorphisms in Candidate Genes for Susceptibility to Bipolar Affective Disorder in the Chromosome 12Q24.31 Region Eric Shink,1 Mario Harvey,1 Monique Tremblay,1 Bernard Gagné,1 Pascal Belleau,1 Catherine Raymond,1 Michel Labbé,1 Marie-Pierre Dubé,2 Ronald G. Lafrenière,2 and Nicholas Barden1* 1 Neuroscience, CHUL Research Center and Laval University, CHUQ Pavillon CHUL, Ste-Foy, Que´bec, Canada Xenon Genetics Research, Inc., Montreal, Canada 2 Previous results from our genetic analyses using pedigrees from a French Canadian population suggested that the interval delimited by markers D12S86 and D12S378 on chromosome 12 was the most probable genomic region to contain a susceptibility gene for affective disorders. Here we present a more detailed genetic analysis of a 7.7 Mb genomic region located on 12q24.31. This region was saturated with 20 microsatellite markers to refine the candidate region and linkage analysis performed in 41 families from the Saguenay-LacSt-Jean (SLSJ) region of Quebec. The results of two point parametric analysis using MFLINK supported the presence of a susceptibility locus on chromosome 12q24.31. Association studies with microsatellite markers using a case/control sample from the same population (n ¼ 401) and analyzed with CLUMP revealed significant allelic associations between the bipolar phenotype and markers NBG6 (P ¼ 0.008) and NBG12 (P < 103). According to these results, we investigated candidate genes in the NBG12 area. We analyzed 32 genes for the presence of polymorphisms in coding sequences and intron/exon junctions and genotyped 22 non-synonymous SNPs in the SLSJ case/control sample. Two uncommon polymorphisms (minor allele frequency 0.03) found in KIAA1595 and FLJ22471 genes, gave P-values below 0.05 with the T1 statistic. Moreover, using haplotype analysis, a nearly significant haplotypic association was observed at the HM74 gene. These results do not give strong support for a role in the susceptibility to bipolar disorder of any of these genes analyzed. However, the significance of rare polymorphisms should be explored by further analyses. ß 2005 Wiley-Liss, Inc. KEY WORDS: bipolar disorder; linkage analysis; association analysis; linkage disequilibrium structure; chromosome 12q24.31 region Grant sponsor: Canadian Institutes of Health Research; Grant sponsor: Xenon Genetics Research, Inc. *Correspondence to: Dr. Nicholas Barden, Neuroscience, CHUQ pavillon CHUL, 2705 Blvd. Laurier, Ste-Foy, Québec, G1V 4G2, Canada. E-mail: email@example.com Received 15 July 2004; Accepted 22 November 2004 DOI 10.1002/ajmg.b.30165 ß 2005 Wiley-Liss, Inc. INTRODUCTION Bipolar affective disorder (BP) is a common complex neuropsychiatric disorder, characterized by the recurrence of depressive and manic/hypomanic episodes, with a worldwide prevalence of 0.4%–1.6%. While evidence for a genetic contribution to BP etiology is supported by family, twin, and adoption studies [Tsuang and Faraone, 1990], the nature of the underlying genetic mechanisms is less clear [Craddock and Jones, 2001]. Despite these restrictions, consistent linkage findings have been reported for some broad chromosomal regions that are likely to harbor deleterious alleles for BP, including 12q23-q24 [Craddock et al., 1994; Dawson et al., 1995b; Ewald et al., 1998, 2002, 2003; Morissette et al., 1999; Degn et al., 2001; Curtis et al., 2003; Ekholm et al., 2003; Shink et al., 2005]. Some candidate genes including Darier’s disease gene (ATP2A2) [Jacobsen et al., 2001b; Jones et al., 2002], PLA2G1B [Dawson et al., 1995a; Jacobsen et al., 1996, 1999], UTL2/CUX2 [Jacobsen et al., 2001a], PAH [Green et al., 2003], LHX5 [Green et al., 2003], DUSP6 [Toyota et al., 2000], and NOS1 [Buttenschon et al., 2004] from chromosomal region 12q23-q24 have already been investigated. Most of these genes are localized in the neighborhood of the Darier’s disease gene. However, no clear association with affective disorder has been reported so far. Our first genome-wide scan suggested the presence of a susceptibility locus on chromosome 12q23-q24 that segregates in two large pedigrees coming from the Saguenay-Lac-St-Jean (SLSJ) area in Quebec [Morissette et al., 1999]. We observed a LOD score value of 2.87 at locus D12S78 under an intraheterogeneity model with recessive segregation mode. Although this population was initially expected to be homogeneous, subsequent analyses did not support the presence of a shared haplotype between most BP subjects [Shink et al., 2003]. More recently, we conducted a second genome-wide scan with 20 families sampled from the SLSJ region and the results again supported the presence of a susceptibility locus for BP on chromosome 12q24 under both parametric and model-free analyses [Shink et al., 2005]. We undertook the characterization of the distal 12q24 genomic region and present here results obtained for a 7.7 Mb interval. We first report linkage-mapping analyses based on 20 highly polymorphic microsatellite markers spanning over this area. We also sought to narrow the subinterval for the screening of candidate genes by performing a case/control study with these markers. After such refinement, we used a candidate gene approach to search for polymorphisms in coding sequences and exon-intron junctions of 32 known genes. Twenty-two non-synonymous single nucleotide polymorphisms (nsSNP) were identified and typed in a case/ control sample from the SLSJ area and analyzed for allelic, genotypic, and haplotypic association with BP affected individuals. The results from this case/control study do not suggest Bipolar Disorder Locus in12Q24.31 that any of these polymorphisms could be a major susceptibility locus for BP. MATERIALS AND METHODS Ascertainment and Diagnosis BP probands selection and family extensions were described elsewhere [Morissette et al., 1999; Shink et al., 2005]. Briefly, diagnosis was established by a panel including at least two psychiatrists and according to the structured clinical interview for DSM-IIIR (SCID I interviews) [Spitzer et al., 1987]. Clarification and possible further interview were requested until agreement was reached in case of any discordance between panel members. For linkage analysis we ascertained 485 individuals from 41 families of the SLSJ region of Quebec. This collection included 21 small pedigrees, totalizing 101 sampled persons never before used for linkage analysis. Diagnoses among the whole genotyped individuals were distributed as follows: 105 bipolar I (BPI) or schizoaffective disorder, bipolar type (Scz), 42 bipolar II (BPII), 54 recurrent unipolar major depression (RUP), and 57 cases of single episode major depression. A case/control sample consisting of 213 individuals from the SLSJ region (BPI (n ¼ 182, mean age at onset 28 11 years [mean SD], 60% female) or BPII (n ¼ 31, mean age at onset 27 11 years [mean SD], 55% female)) was established for association analysis. Although most of the case subjects had a family history of affective disorders, they were all unrelated within immediate generations. Eleven spouses diagnosed as normal and without child with major affective disorder diagnosis were included in the control group. The remainder of controls subjects (n ¼ 177), all from the SLSJ region, were recruited from non-psychiatric genetic projects, since there is no significant gain of power to screen controls for psychiatric disorders with a lifetime risk of about 1% [Owen et al., 1997] and, as far as we could ascertain were representative of the general population. No attempt was made to match the controls and cases by age. Chi-squared test on 2 2 contingency table displayed no significant difference between distributions of gender in the case and control groups (P ¼ 0.318). 51 Canada). Amplification reactions were done using the Taq platinum DNA polymerase (Invitrogen). The amplification products were purified on GF/C 384-well plates from Whatmann (VWR, Montreal, Canada). PCR products (30 ml) using the PB buffer from Qiagen as binding buffer. The purified samples were quantified with the PicoGreen dsDNA quantitation assay (Invitrogen) and sequenced from both strands using the DYEnamic ET terminator cycle sequencing kit (Amersham Biosciences, Baie D’Urfé, Canada). We slightly modified the manufacturer’s protocol. Briefly, in a sequencing reaction we added: 4 ng of DNA, 1 pmole of sequencing primer, 0.2 ml of ET sequencing mix, 1.8 ml of dilution buffer (65 mM Tris, 16.3 mM MgCl2, 12.5% glycerol, pH ¼ 9.5), and completed to 5 ml with water. We ran the following cycling program for 20– 25 cycles: 958C, 20 sec; 598C, 60 sec. The post-sequencing clean up was done by ethanol/salt precipitation. The sequencing products were dissolved in 7 ml of HIDI formamide (Applied Biosystems, Inc, LA), and were resolved on an ABI PRISM 3730XL DNA analyzer (Applied Biosystems, Inc., LA). The sequencing data were analyzed for the presence of polymorphisms by using the STADEN package [Bonfield et al., 2002]. Denaturing HPLC (DHPLC) Method PCR primers were designed using PrimerSelect (DNASTAR, Madison, WI) and purchased from BioCorp (Montreal, Canada). Samples were amplified using Taq polymerase (Qiagen). All amplified fragments were tested by agarose gel electrophoresis, then pooled 2X, denatured by heating to 948C for 5 min, then reannealed slowly by a controlled cooling to 258C over the course of 1.5 hr in a PE 9700 thermocycler. Before DHPLC analysis of PCR products, amplicon sequences were used to determine denaturation temperatures and buffer gradients with the WAVEMAKER software for each amplicon. Samples were then injected onto a DNASepHT column in a model 3500HT WAVE dHPLC apparatus running WAVEMAKER 4.1 software (Transgenomic, Inc., Omaha, Nebraska). Elution profiles were visually inspected and grouped into ‘‘types,’’ such as T1 and T2. Then, results were compiled into an Excel worksheet for further processing. The amplification products showing DHPLC variants were sequenced as described earlier to determine the variations. Genomic DNA Preparation Blood samples from each individual were collected in 10-ml K3 EDTA Vacutainer tube (Becton–Dickinson) and genomic DNA was isolated by Puregene DNA Isolation kit (Gentra Systems) according to the manufacturer’s instructions. DNA was solubilized in 500 ml of DNA Hydration Solution and the final concentration was adjusted to 300–400 mg/ml by spectroscopy at 260 nm. Mutation Analysis We searched for single nucleotide polymorphisms and other variations in coding sequences and exon–intron boundaries of genes using direct sequencing and dHPLC strategies. The starting sample was composed of 16 unrelated BP affected individuals selected to maximize the number of different haplotypes and also according to their link with families that gave positive genetic linkages on chromosome 12q24. The power to identify a diallelic polymorphism with this sample was established at 80% assuming a minor allele frequency of 5% with simple binomial sampling statistics. All novel SNPs identified have been submitted to the NCBI dbSNP. Sequencing Method The targeted sequences were amplified by PCR. Amplification primers were purchased from Invitrogen (Burlington, Genotyping of Microsatellite Markers We used a fluorescent-based method for the genotyping of microsatellite markers. Briefly, the region encompassing each repeated sequence was amplified by PCR using an unlabeled primer and a fluorescent-labeled primer (Applied Biosystems, Inc, CA). The marker-associated dyes and the corresponding amplification product length are listed in Table I. The amplification reaction is performed using 50 ng of DNA sample, 0.2 unit of Taq platinum DNA polymerase (Invitrogen, Burlington, Canada), 20 mM Tris-Cl (pH 8.4), 50 mM KCl, 1.5 mM MgCl2, 100 mM of each dNTP, and 1.5 mM of each primer in a final volume of 7 ml. The samples are incubated at 958C for 3 min to activate the Taq platinum DNA polymerase, then 10 cycles of PCR amplification were done as follow: 958C, 15 sec; 588C, 15 sec, 728C, 30 sec; after that 15 cycles as follow: 898C, 15 sec; 588C, 15 sec, 728C, 30 sec. Finally, the samples are incubated at 728C for 30 min. Then, they were pooled according to their dye-labeled primer and their PCR product length (pool of four samples), resolved on an ABI 3100 DNA analyzer (Applied Biosystems, Inc, CA). The raw data were analyzed using Genemapper2 (Applied Biosystems, Inc, CA), and compiled in a 4 D database (ACIUS) designed in a Macintosh environment as previously described [Morissette, 1992]. Allelic frequencies were estimated with VITESSE [O’Connell and Weeks, 1995] from the pedigree sample [Boehnke, 1991]. 52 Shink et al. TABLE I. List of Microsatellite Markers Used for Linkage Analyses and Association Studies Dyea Locus Allele length (bp) Het(%) Position (bp) NBG_11 VIC 204–218 66 119963576 D12S1666 NBG_5 6FAM VIC 241–281 253–261 67 38 120066644 120114441 D12S1721 NBG_8 VIC VIC 263–299 166–188 72 73 120469198 121097458 NBG_6 NED 182–218 74 121396927 NBG_9 VIC 156–180 69 121695884 NBG_10 6FAM 174–186 50 121976850 D12S1349 NBG_12 NED NED 172–196 165–207 79 64 122090292 122470449 NBG_4 NED 171–199 67 124214016 NBG_3 VIC 182–206 65 124327585 D12S378 NBG_2 VIC VIC 152–176 128–160 73 54 124387047 124923002 D12S1614 D12S342 D12S340 D12S1639 D12S1634 D12S2075 NED VIC PET 6FAM PET 6FAM 167–177 218–240 242–250 221–255 151–169 275–311 72 74 62 77 51 71 125599529 125607947 125865825 126102199 126888404 127685134 Primersb TCATTTGTCCTATTCCCATTCTG(F) GTTTCTT-TGTAGAGTTTTCTGAATGTCCGC(R) UniSTS:5781 TAAGCCAGCTGATTCCCAAG(F) GTTTCTT-CTGGACAATATAGCGAGGCT(R) UniSTS:16176 TCACGTCCTTCTTTCTTTGC(F) GTTTCTT-GCATGCAGCTAAAAAGTGCT(R) CGAACCAGCCCTAACCTAGC(F) GTTTCTT-GGACAGGATGTCGTGGGAT(R) TTGTGTGTCCATTTTTGACCA(F) GTTTCTT-TGACGAAATTGCCTAATGATG(R) GTGACCCTGGGTTCCAAGCA(F) GTTTCTT-GACCTGGGAGGTCAGGG(R) UniSTS:31135 GGTGAGAGAGCAAGATCCTTTCTATT(F) GTTTCTT-AGATGAGAACCCTAGCCAGCC(R) GTTTGTTACTTCACATTTTGACTGG(F) GTTTCTT-TCTTGGAACTTCAGGCCAAC(R) ATGTACCCAGGAAACCAGCA(F) GTTTCTT-GTTTGACTGTTGCAATCATGTTACT(R) UniSTS:48801 CTGGGCAAGAGAGTGAAACC(F) GTTTCTT-TGAACTCCACCAGGGTTTGT(R) UniSTS:66768 UniSTS:34698 UniSTS:70202 UniSTS:14122 UniSTS:62030 UniSTS:59386 The markers were from public databases (D12SXXXX) or our laboratory (NBG_XX). Novel microsatellite markers (internal code NBG) have been submitted to the public database (NCBI UniSTS). For each marker, we indicate the fluorescent dye used to label the forward primer; the observed amplification product size; the allele heterogeneity; the position on chromosome 12 (at the end of repeats) based on build 114 of the human genome assembly; and either primer sequences or marker references in the UniSTS database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db¼unists). All markers contain CA repeats except for D12S378 and D12S2075, which are tetranucleotide repeats. a The fluorescent dyes are trademarks of PE Corporation. b GTTTCTT, pig-tail sequence. Genotyping of SNPs SNPs were genotyped by direct sequencing of the amplification products. The sequencing traces for each individual were automatically typed for the corresponding SNP using a homedeveloped program, GENO.pl and results of SNP genotyping were compiled in the 4 D database cited above. SNP allelic frequencies were computed directly from the control group and deviation from Hardy–Weinberg equilibrium (HWE) was assessed with the exact test from Genepop software package (http://wbiomed.curtin.edu.au/genepop). Microsatellite Marker Statistical Analysis Two hierarchical affection status models (ASM) were used to define the affected phenotype for linkage analysis since BP are made up of a ‘‘spectrum’’ of associated disorders that could be modulated by common susceptibility genes [Tsuang and Faraone, 1990; Berrettini, 1997]. The first model (ASMI) classified BPI, Scz and BPII individuals as affected, and single or recurrent episode major depression diagnosis as phenotype unknown. This last phenotype was extended to include subjects whose status was unclear. All other diagnoses were considered as unaffected. The second model (ASMII) was similar to ASMI but the affected category was broadened to include individuals with recurrent episode major depression. The MFLINK software package (version 2.0) was used for linkage analysis with microsatellite markers [Curtis and Sham, 1995; Curtis et al., 1999; Sham et al., 2000]. This software maximizes the LOD score over a set of disease-model parameters subject to constraints (MLOD) such that the statistic 4.6xMLOD has an asymptotic two-tailed chi-squared distribution with one degree of freedom. The disease prevalence was set to 1% and the mode of inheritance varied from recessive to dominant during the process of maximization. MLODs were computed for multiplex pedigrees and for their nuclear components since the sample contained large densely affected kindred, which are more likely to characterize the segregation of multiple susceptibility genes [Durner et al., 1992]. No correction was applied to take into account the presence of two ASMs and the use of the whole pedigree sample and its nuclear components. Indeed, there was no clear way to correct for multiple testing since these tests were correlated. The association studies were done with our dense set of microsatellite markers to reduce the region of interest [Weeks and Lathrop, 1995; Hardy and Singleton, 2000]. The association hypothesis, which assumes correlated occurrence of DNA sequence and bipolar phenotype on a population scale, was tested with the software package CLUMP Version 1.6 [Sham and Curtis, 1995]. The usual chi-squared statistic on the raw contingency table (T1 statistic) and the largest chi-squared statistic calculated by comparing one column of the original table against the total of the other columns (T3 statistic) were utilized. One thousand simulations were done to establish empirical P-value with CLUMP. Bipolar Disorder Locus in12Q24.31 Haplotyping was done on the pedigrees with the GENEHUNTER software package [Kruglyak et al., 1996]. SNPs Statistical Analysis As for microsatellite markers, the association hypothesis with SNPs was tested with CLUMP. Only the T1 statistic was used for allelic association, while both T1 and T3 statistics were used to test genotypic association. SNP haplotypes were assessed within genes with the software package PHASE Version 1.0 [Stephens et al., 2001]. Considering the predicted haplotypes as alleles, T1 and T3 statistics were computed for allelic and genotypic association analysis. No correction for multiple testing was applied for association analysis since this work was prospective. RESULTS Linkage and Association Results With Microsatellite Markers The linkage results from the whole pedigree series and their nuclear components are shown in Figure 1 for ASMI and ASMII. Most maximum MLOD scores were observed when pedigrees were broken into nuclear families (15/20). All microsatellite markers show at least one MLOD value with P-values less than 0.05 except at D12S1634. The maximum MLOD value of 3.85 (P ¼ 0.00003) was reached at D12S1721 under ASMI and pedigrees broken into branches. Under ASMII and broken pedigrees, its MLOD value decreased to 3.18 (P ¼ 0.00013) but two close markers, NBG11 and NBG6, gave MLOD values greater or equal to 3.28 with associated P-values less or equal to 0.0001. No correction for multiple testing was applied because of correlated tests. NBG3 was the most distal marker with maximum MLOD value exceeding 53 the threshold value of 3. Maximum MLOD values with jointed P-values of 0.00154 and 0.00444 were reported respectively for markers D12S342 and D12S1639 previously reported by others groups as described above. HWE hypothesis was satisfied at the 5% level for each microsatellite marker after application of the Bonferroni corrections for multiple testing. Table II lists empirical P-values observed with CLUMP for allele and genotype association analysis. Empirical P-values less than 0.005 were observed at marker NBG12 for T1 and T3 statistics under allelic association analysis. T1 statistic suggested allelic association between BP and NBG6 (empirical P ¼ 0.008). Moreover, a borderline significant empirical P-value of 0.023 was observed at the most distal marker D12S2075. Based on these results and also on previous linkage studies, the candidate region on chromosome 12q24.31 must include sequences between NBG11 and D12S2075. However, the distal region from the associated marker NBG12 was chosen as the best to begin our mutation screening on the basis of an extensive haplotype analysis (results not shown). Mutation Screening in Candidate Genes We undertook a large mutation analysis on almost all known genes located between NBG12 and D12S342. This 2.5 Mb genomic region, of which 38 known genes have been annotated, is included in the contig NT_09755.15 from the NCBI. In this case, known genes are defined as protein-coding genes supported by at least one experimental mRNA, even though function may be known or unknown. We selected 32 of these 38 genes to be screened for the presence of sequence variations, mostly in coding regions (see Table III). When known, the 50 and 30 -untranslated sequences were also partially or totally analyzed. We did not analyze any promoter region. The Fig. 1. Two-point MLOD observed for the whole pedigree series and their nuclear components using MFLINK. Analysis was conducted by using ASMI (*–* for whole pedigree; &- - -&for nuclear components) and ASMII (~–~for whole pedigree; ^- -^for nuclear components). 54 Shink et al. TABLE II. Empirical P-values Observed With CLUMP for Statistics T1 and T3 for Allelic and Genotypic Analyses of Microsatellite Markers (1000 Simulation Replicates) Effective Alleles Genotypes Locus Case Control T1 (P-value) T3 (P-value) T1 (P-value) T3 (P-value) NBG11 D12S1666 NBG5 D12S1721 NBG8 NBG6 NBG9 NBG10 D12S1349 NBG12 NBG4 NBG3 D12S378 NBG2 D12S1614 D12S342 D12S340 D12S1639 D12S1634 D12S2075 204 208 213 210 213 213 213 213 212 213 207 209 211 210 210 211 209 209 211 203 98 175 179 176 179 179 175 175 180 175 178 175 180 170 179 180 179 180 181 181 0.250 0.366 0.969 0.693 0.754 0.008 0.759 0.521 0.887 0.002 0.418 0.171 0.171 0.896 0.803 0.394 0.890 0.087 0.361 0.023 0.421 0.543 0.934 0.463 0.921 0.356 0.768 0.178 0.864 <103 0.506 0.829 0.405 0.749 0.692 0.740 0.869 0.170 0.248 0.157 0.680 0.393 0.997 0.805 0.973 0.172 0.690 0.122 0.782 0.018 0.813 0.601 0.540 0.210 0.710 0.445 0.895 0.652 0.505 0.085 0.553 0.476 1.000 0.838 0.929 0.449 0.606 0.173 0.816 0.552 0.545 0.897 0.560 0.613 0.831 0.622 0.838 0.295 0.590 0.451 TABLE III. List of Genes Analyzed for the Presence of Mutations Locus B3GNT4 SMAC VPS33A RSN FLJ14796 FLJ11021 KNTC1 HM74a HM74 GPR81 DENR HIP1R FLJ12750 ABCB9 FLJ13491 ARL6IP4/SRP25 PITPNM2/NIR3 MPHOSPH9 FLJ22255/FLJ38663 CDK2AP1 SBNO1 MGC7036/FLJ32372 RNP24 KIAA1595/DDX55 EIF2B1 GTF2H3 FLJ12975 ATP6V0A2/TJ6 KIAA2017/DNAH10 FLJ22471 DKFZp761B128 SMRT/N-CoR2 RefSeq Orientation Non-coding and coding exons UTR Coding length (bp) NM_030765.1 NM_019887.2 NM_022916.2 NM_002956.1 AK027702.1 NM_023012.2 NM_014708.2 NM_177551.2 NM_006018.1 NM_032554.2 NM_003677.2 XM_290592.1 NM_024667.1 NM_019625.1 NM_024623.1 NM_016638.1 NM_020845.1 NM_022782.1 NM_152269.1 NM_004642.2 AB014772.1 NM_145058.1 NM_006815.2 BC030020.1 NM_001414.1 NM_001516.3 NM_024809.2 NM_012463.1 XM_113706 NM_025140.1 NM_152437.1 NM_006312.1 þ þ þ þ þ þ þ þ þ þ þ þ þ þ 2 6 13 24 14 10 (1)63 1 1 1(1) (1)7 32 4 (1)11 (3)6 5 24 (1)16(2) (1)2 4 31* 4 4 14 8 13 18 20 7* (1)4 (4)1 47 T T T T T 3P T T T 3P 3P T 3P T T T 3P T T T 5NA 3P T T T 3NA T 3NA T 5NA T T 1,137 720 1,791 4,284 2,124 1,305 6,630 1,092 1,164 1,041 598 3,207 858 2,172 873 657 4,050 2,079 501 348 4,179 636 606 1,803 918 927 2,094 2,571 1,155 996 786 7,554 The genes are listed from centromere to telomere. The column orientation indicates the strand corresponding to mRNA sequences. The number (in brackets) at left or at center of coding exons specifies the number of 50 or 30 untranslated exons, respectively. The asterisks (*) designate the sequence partially annotated into genomic contig. The column UTR describes the status for the analysis of untranslated regions; T, totally analyzed; 3P, 30 UTR partially analyzed; 5NA, 50 UTR not analyzed; 3NA, 30 UTR not analyzed. Bipolar Disorder Locus in12Q24.31 remaining six genes were eliminated according to these features: The incompleteness of the genomic annotation (MONDOA), the function and expression (TSP-NY), overlapping with well-known locus (FLJ213769), not annotated when this study was started (MGC35140, Loc347945, and Loc144347). The sequences to be screened were dictated by the mRNA Reference Sequence (RefSeq) from the NCBI, except for FLJ14796, SBNO1, SMRT/N-CoR2, and KIAA1595 of which the RefSeq did not correctly represent the most probable full length coding sequences. Moreover, EST clustering analyses (http://genenest.molgen.mpg.de/) suggested that some genes (ABCB9, CDK2AP1, EIF2B1, FLJ11021, FLJ13491, FLJ14796, FLJ22471, MPHOSPH9, SRp25, and SMRT/N-CoR2) could generate different protein variants by alternative exon usage. As already mentioned, the RefSeq associated to these loci corresponds to the most probable transcripts (Cluster having the largest number of associated mRNA and EST), but we screened for exons generating all possibilities of transcription variants. Among listed genes, we could point out some of the most interesting candidates based on their known or suspected function. HM74 and GPR81 belong to the G protein-coupled receptor superfamily. This type of receptor is involved in regulation of intracellular second messengers, such as cAMP, inositol phosphates, diacylglycerol, and calcium ions [Karasinska et al., 2003]. Interestingly, the HM74 locus contains a duplicate genomic region, coding for two HM74 variants, HM74 and HM74a, which are 94% identical. The identification of mRNA and ESTs associated to each form suggest that both variants could be expressed. Recently, nicotinic acid has been identified as a ligand for HM74 and HM74a in adipocytes [Wise et al., 2003]. On the other hand, GPR81 remains an orphan receptor. NIR3 is also another interesting candidate gene based on its expression in the dentate gyrus and its potential involvement in phosphoinositol signaling. Lithium is known to directly inhibit the enzyme monophosphate phosphatase implicated in phosphoinositide metabolism [Lenox and Wang, 2003]. In this mutation analysis we identified 169 polymorphisms, including 13 in 50 -untranslated sequences (50 -UTR), 78 in introns, 28 in 30 -UTR, and 50 in coding regions (data not shown). Among these coding SNPs, 25 (50%) are non-synonymous polymorphisms (nsSNPs) and we found one glutamine repeat (CAG12-16) in the SMRT/N-CoR2 gene. The latter and other point mutations in SMRT/N-CoR2 gene led us to independently investigate this locus in more detail [Shink et al., 2005]. All other missense mutations, listed in Table IV, have been genotyped and analyzed for association with BP. SNP Association Results All non-synonymous SNPs identified in our mutation screening were typed in a case/control sample (401 individuals) from the SLSJ region. Allelic and genotypic association results are shown in Table V. Considering the power of our study, we selected to maximize the number of positive associations (type I error) by setting up a threshold for significant results at P-values ¼ 0.05 without correction for multiple testing and to replicate positive results in an independent sample. We found 15 of 22 SNPs had minor allele frequencies of less than 20%. Empirical P-value was less than 0.05 for T1 statistic under genotypic analysis at SNPs GC29E06A and GC32E04A located within the KIAA1595/DDX55 and the FLJ22471 loci respectively. The SNP GC29E06A describes the amino acid change Glu154Lys. The DDX55 protein is a member of the DEAD-box family, which is mostly involved in ribosome biogenesis and translation initiation [Abdelhaleem et al., 2003]. The glutamic acid 154 residue is conserved between three DDX55 orthologs (human, rat, and mouse), and is replaced by the other negatively charged amino acids, aspartic acid, in the Tetraodon homolog (data not shown). This residue is included in the DEAD box domain, but is not conserved among DEAD boxcontaining proteins. Evidence for a susceptibility gene was observed at SNP GC32E04A with an empirical P-values of 0.001, but the control sample did not satisfy HWE even after correction for multiple testing (uncorrected P < 0.001) since all TABLE IV. List of Genotyped Missense Variations Locus SNP RSN GC04E19A GC04E14A GC04E12A GC06E04A GC07E08A GC07E25A GC07E57A GC09E01B GC09E01A GC10E01E GC10E01C GC10E01D GC10E01B GC10E01A GC18E01A GC22E16A GC29E04A GC29E06A GC29E08A GC24E19A GC32E04A GC32E03A FLJ11021 KNTC1 HM74v02 HM74v01 NIR3 SBNO1 KIAA1595 TJ6 FLJ22471 55 Alleles A-C (A1178S) G-C (E1034D) T-C (S895P) A-G (H88R) G-T (K245N) G-C (E738D) T-G (V2021G) G-A (M317I) C-T (R311C) G-C (M346I) G-A (M317I) G-A (R253H) C-T (L198F) C-A (P173T) C-T (P9L) G-A (S728N) G-C (V101L) G-A (E154K) A-G (N264S) C-T (A813V) G-A (R281H) A-T (S70C) Genomic position (bp) dbSNP 122572096 122620929 122626006 122811947 122839214 122866189 122911347 122995306 122995326 123008673 123008760 123008953 123009119 123009194 123327538 123614645 123815976 123817228 123821709 123965449 124145702 124151249 — rs1129167 — — — — — — — rs1696351 rs2454727 rs2454726 rs676823 rs1798192 — rs1060105 — — — — — rs11057401 The chromosomal locations are dictated by the contig assembly build 33 from the NCBI. The column ‘Alleles’ describes the punctual nucleotide variation (WT-mutant), and the corresponding amino acid modification. The wild type sequences correspond to RefSeq in Table III. Novel SNPs have been submitted to the NCBI dbSNP and the reference SNP ID is given for SNPs that have already been characterized. 56 Shink et al. TABLE V. Empirical P-Values Observed With CLUMP for Allelic and Genotypic Analyses of Coding SNPs (1000 Simulation Replicates) Effective Gene (a; b)* RSN (5; 2) FLJ11021 KNTC1 (5; 1) HM74a (4; 2) HM74 (7; 2) NIR3 SBNO1 KIAA1595 (4; 1) TJ6 FLJ22471 (4; 2) Alleles Genotypes Locus Allele frequencies Case Control T1 (P-value) T1 (P-value) T3 (P-value) GC04E19A GC04E14A GC04E12A GC6E04A GC07E08A GC07E25A GC07E57A GC09E01B GC09E01A GC10E01E GC10E01C GC10E01D GC10E01B GC10E01A GC18E01A GC22E16A GC29E04A GC29E06A GC29E08A GC24E19A GC32E04A GC32E03A A (0.018); C (0.982 C (0.646); G (0.354) A (0.879); G (0.121) A (0.997); G (0.003) G (0.906); T (0.094) C (0.972); G (0.028) G (0.088); T (0.912) C (0.679); T (0.321) A (0.118); G (0.882) C (0.062); G (0.938) A (0.347); G (0.653) A (0.446); G (0.554) C (0.523); T (0.477) A (0.463); C (0.537) A (0.045); G (0.955) A (0.204); G (0.796) C (0.020); G (0.980) A (0.022); G (0.978) A (0.898); G (0.102) A (0.012); G (0.988) A (0.030); G (0,970) A (0,607); T (0,393) 201 205 203 196 204 201 200 204 204 202 202 200 200 200 203 192 205 200 205 204 193 205 169 171 170 159 170 162 170 179 178 178 177 175 177 178 177 164 175 160 176 166 167 169 0.325 0.776 0.845 1.000 1.000 0.806 1.000 0.350 0.289 0.520 0.079 0.714 0.240 0.644 0.230 0.779 0.572 0.129 0.178 0.760 1.000 0.228 0.319 0.071 0.652 1.000 0.251 0.835 0.097 0.645 0.382 0.533 0.213 0.785 0.148 0.660 0.235 0.908 0.587 0.027 0.393 0.754 0.001 0.365 0.319 0.073 0.621 1.000 0.462 0.835 0.363 0.631 0.207 0.383 0.215 0.781 0.168 0.659 0.235 0.888 0.587 0.078 0.354 0.754 0.287 0.385 *a, maximum number of haplotypes estimated by PHASE over the case and the control sample for the gene; b, number of major haplotype that accounted for an estimated 80% of the observed chromosomes in both sample. heterozygous individuals (10) had BPI or BPII diagnosis. We verified the typing for the SNP GC32E04A to eliminate the possibility of experimental errors. A slight increase of heterozygous genotype was observed at GC29E06A in the control group (4%) versus the case group (0.5%). Table V shows the maximum number of haplotypes estimated by PHASE over the case and the control sample for each gene with more than one genotyped SNP. It appeared that one or two major haplotypes accounted for an estimated 80% of the observed chromosomes in each sample. Case/control studies of SNP haplotypes led to an empirical P-value of 0.030 for T1 statistic with the gene HM74. PHASE allowed us to reconstruct only 7 of 32 possible haplotypes within HM74, the largest difference between case and control haplotype frequencies was an increase of 4% in the case group. Moreover, a haplotype found in the controls with a frequency of 2% was not observed in cases. An empirical P-value of 0.020 was obtained with T1 statistic when the genotypes of haplotypes distribution were compared between cases and controls at gene FLJ22471. No other association was significant at the 5% level, including KIA1595. DISCUSSION These results present the first exhaustive candidate genes analysis distal to the Darier’s disease gene locus, and point out possible deleterious alleles in, or close to, three genes that could have a small effect on the incidence of BP. We first showed a linkage analysis over 7.7 Mb, using 20 microsatellite markers from NBG11 to D12S2075. Most maximum MLOD values (15/20) have been shown with nuclear families that could be explained by intra-family heterogeneity. This result is coherent with our previous report suggesting that different ancestors have introduced one or more common deleterious alleles [Shink et al., 2003]. In the same way, the perspective that a susceptibility gene to BP can segregate in many different ethnic groups argues in favor of a common deleterious allele. However, as mentioned elsewhere [Curtis et al., 2003], we may not exclude the possibility of a rare allele-causing disease since this genomic region has been highlighted in very large pedigrees. The highest two-point MLOD scores, greater or equal to 3.28 (P-value 0.0001), were observed at markers NBG11, D12S1721, and NBG6, which are more proximal than those previously reported. For many reasons, the replication of significant linkage results is not a foregone conclusion [Suarez et al., 1994; Altmuller et al., 2001] and some positive peaks may drift over 15 cM from one study to another [Hauser et al., 1996; Roberts et al., 1999]. Consequently, the presence of maximum MLOD values displayed at several distant markers may reflect a common susceptibility gene that segregates in many different populations, such as Danish, English, Faroese, French Canadian, and Icelandic [Ewald et al., 1998; Degn et al., 2001; Curtis et al., 2003; Ekholm et al., 2003]. In fact, the hypothesis that a susceptibility gene to BP is common to Faroe Islands and SLSJ populations is supported by results observed at D12S2075 by Degn et al.  and this report. Accordingly, the recent genome-wide scan by Curtis et al.  supports linkage at marker D12S342. Unfortunately, D12S342 was their unique locus that overlapped with our studied markers and it will be interesting to see a more detailed linkage analysis surrounding this marker. Obviously, it is possible that the 12q23-24 region could contain more than one locus involved in BP. The analysis of the genomic region surrounding the positively associated marker NBG12 was based on a candidate gene approach and prioritized non-synonymous SNPs for case/control studies. This strategy is mainly supported by the hypothesis that such polymorphisms are likely to affect the risk of an affected phenotype, and by mutational studies of Mendelian diseases indicating that missense/nonsense polymorphisms are the most frequent disease-causing mutations (59%) [Tabor et al., 2002; Botstein and Risch, 2003]. By this approach we also wanted to maximize the chance of finding a biologically relevant association. The association studies with 22 non-synonymous mutations revealed significant genotypic associations at the 5% level for two uncommon variations located in genes KIAA1595 and FLJ22471, which encode for Bipolar Disorder Locus in12Q24.31 DDX55 and limkain beta 2 proteins, respectively. The distance between SNPs is about 328 Kb and they did not show LD (data not shown). The exact function of DDX55 is unknown, but as mentioned earlier, it belongs to the DEAD-box family of proteins involved in ribosome biogenesis and translation initiation. The yeast homolog, Spb4p, plays an important role in maturation process of the 25S and 5.8S ribosomal RNA [de la Cruz et al., 1998]. Moreover, Spb4p is a RNA helicase essential for cell viability [Sachs and Davis, 1990]. Considering this, it is unlikely that KIAA1595/DDX55 would be associated to affective disorders. The most striking result was observed for SNP GC32E04A in Limkain beta 2 gene where all heterozygous individuals were affected. The typing results for this polymorphism deviate from the HW equilibrium and we excluded any cause by experimental errors. Accordingly, this deviation could be due to a chance sampling error or, on the other hand, it could be interpreted as the presence of a diseaseassociated genotype. Considering the latter hypothesis, the low frequency of this polymorphism and the BP prevalence of about 1%, it is not likely that we would observe any heterozygote in the control sample. The contribution of heterozygous mutations in disease susceptibility has already been reported [Morissette et al., 1998]. For the limkain beta 2 proteins, we noted a coil-coiled motif that could permit either selfinteractions or with other cytoskeleton-associated proteins, thus we could postulate that diverging alleles would generate a heterogeneous pool of proteins, which improperly interact one with another. Alternatively, the heterozygotes at locus GC32E04A could potentiate the effect of preliminary mutated alleles as already described in recent reports about heterozygous mutations on Bardet–Biedl syndrome [Badano et al., 2003]. However, since this represents only 5% (10/193) of bipolar subjects this could not account for a majority of BP cases in our sample, and no specific BP sub-phenotype has been associated to this genotype. This finding has to be replicated in an independent sample. In conclusion, the data reported here present a major fine mapping effort over the chromosomal 12q24.31 area. Our association studies were focused on a narrow portion of this genomic region, motivated by a positive allelic association with the microsatellite marker NBG12. We analyzed 32 genes for mutation detection, prioritized nsSNPs for genotyping and found polymorphisms associated with bipolar disorder in three genes. While we do not believe that any of them could be responsible for the linkage peak observed in our family studies, they could play a role in the promotion of BP severity, or they could be in LD with more frequent disorder-causing alleles. Finally, while this study shows the necessity to consider candidate genes around D12S2075 since it has been implicated in two independent studies, the positively associated region upstream of NBG12 cannot be ignored. ACKNOWLEDGMENTS The authors thank patients and their families for participating in the study as well as Dr. Laberge, Dr. Raymond, Dr. Brown, and Dr. M. Morissette for control samples. REFERENCES Abdelhaleem M, Maltais L, Wain H. 2003. The human DDX and DHX gene families of putative RNA helicases. Genomics 81:618–622. Altmuller J, Palmer LJ, Fischer G, Scherb M, Wjst M. 2001. Genomewide scans of complex human diseases: True linkage is hard to find. Am J Hum Genet 69:936–950. Badano JL, Kim JC, Hoskins BE, Lewis RA, Ansley SJ, Cutter DJ, Castellan C, Beales PL, Leroux MR, Katsanis N. 2003. Heterozygous mutations in BBS1, BBS2 and BBS6 have a potential epistatic effect on Bardet-Biedl patients with two mutations at a second BBS locus. Hum Mol Genet 12:1651–1659. 57 Berrettini W. 1997. Molecular linkage studies of manic-depressive illness. In: Blum K, Noble EP, editors. Handbook of psychiatric genetics. New York: CRC Press, Inc. pp 261–272. Boehnke M. 1991. Allele frequency estimation from data on relatives. Am J Hum Genet 48:22–25. Bonfield J, Beal K, Cheng Y, Jordan M, Staden R. 2002. The package Staden 2002.0. Cambridge, UK: The MRC laboratory of molecular biology. Botstein D, Risch N. 2003. Discovering genotypes underlying human phenotypes: Past successes for mendelian disease, future approaches for complex disease. Nat Genet 33(Suppl):228–237. Buttenschon HN, Mors O, Ewald H, McQuillin A, Kalsi G, Lawrence J, Gurling H, Kruse TA. 2004. No association between a neuronal nitric oxide synthase (NOS1) gene polymorphism on chromosome 12q24 and bipolar disorder. Am J Med Genet 124B:73–75. Craddock N, Jones I. 2001. Molecular genetics of bipolar disorder. Br J Psychiatry 178:128–133. Craddock N, Owen M, Burge S, Kurian B, Thomas P, McGuffin P. 1994. Familial cosegregation of major affective disorder and Darier’s disease (keratosis follicularis). Br J Psychiatry 164:355–358. Curtis D, Sham PC. 1995. Model-free linkage analysis using likelihoods. Am J Hum Genet 57:703–716. Curtis D, Zhao JH, Sham PC. 1999. Comparison of GENEHUNTER and MFLINK for analysis of COGA linkage data. Genet Epidemiol 17:S115– S120. Curtis D, Kalsi G, Brynjolfsson J, McInnis M, O’Neill J, Smyth C, Moloney E, Murphy P, McQuillin A, Petursson H, Gurling H. 2003. Genome scan of pedigrees multiply affected with bipolar disorder provides further support for the presence of a susceptibility locus on chromosome 12q23-q24, and suggests the presence of additional loci on 1p and 1q. Psychiatr Genet 13:77–84. Dawson E, Gill M, Curtis D, Castle D, Hunt N, Murray R, Powell J. 1995a. Genetic association between alleles of pancreatic phospholipase A2 gene and bipolar affective disorder. Psychiatr Genet 5:177–180. Dawson E, Parfitt E, Roberts Q, Daniels J, Lim L, Sham P, Nothen M, Propping P, Lanczik M, Maier W, Reuner U, Weissenbach J, Gill M, Powell J, McGuffin P, Owen M, Craddock N. 1995b. Linkage studies of bipolar disorder in the region of the Darier’s disease gene on chromosome 12q23-24.1. Am J Med Genet 60:94–102. de la Cruz J, Kressler D, Rojo M, Tollervey D, Linder P. 1998. Spb4p, an essential putative RNA helicase, is required for a late step in the assembly of 60S ribosomal subunits in Saccharomyces cerevisiae. RNA 4:1268–1281. Degn B, Lundorf MD, Wang A, Vang M, Mors O, Kruse TA, Ewald H. 2001. Further evidence for a bipolar risk gene on chromosome 12q24 suggested by investigation of haplotype sharing and allelic association in patients from the Faroe Islands. Mol Psychiatry 6:450–455. Durner M, Greenberg DA, Hodge SE. 1992. Inter- and intrafamilial heterogeneity: Effective sampling strategies and comparison of analysis methods. Am J Hum Genet 51:859–870. 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 12:1907–1915. Ewald H, Degn B, Mors O, Kruse TA. 1998. Significant linkage between bipolar affective disorder and chromosome 12q24. Psychiatr Genet 8: 131–140. Ewald H, Flint T, Kruse TA, Mors O. 2002. A genome-wide scan shows significant linkage between bipolar disorder and chromosome 12q24.3 and suggestive linkage to chromosomes 1p22-21, 4p16, 6q14-22, 10q26 and 16p13.3. Mol Psychiatry 7:734–744. 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 120B:63–71. Green EK, Elvidge GP, Owen MJ, Craddock N. 2003. Mutational analysis of two positional candidate susceptibility genes for bipolar disorder on chromosome 12q23-q24: Phenylalanine hydroxylase and human LIMhomeobox LHX5. Psychiatr Genet 13:97–101. Hardy J, Singleton A. 2000. The future of genetic analysis of neurological disorders. Neurobiol Dis 7:65–69. Hauser ER, Boehnke M, Guo SW, Risch N. 1996. Affected-sib-pair interval mapping and exclusion for complex genetic traits: Sampling considerations. Genet Epidemiol 13:117–137. 58 Shink et al. Jacobsen N, Daniels J, Moorhead S, Harrison D, Feldman E, McGuffin P, Owen MJ, Craddock N. 1996. Association study of bipolar disorder at the phospholipase A2 gene (PLA2A) in the Darier’s disease (DAR) region of chromosome 12q23-q24.1. Psychiatr Genet 6:195–199. Jacobsen NJ, Franks EK, Owen MJ, Craddock NJ. 1999. Mutational analysis of phospholipase A2A: A positional candidate susceptibility gene for bipolar disorder. Mol Psychiatry 4:274–279. Jacobsen NJ, Elvidge G, Franks EK, O’Donovan MC, Craddock N, Owen MJ. 2001a. CUX2, a potential regulator of NCAM expression: Genomic characterization and analysis as a positional candidate susceptibility gene for bipolar disorder. Am J Med Genet 105:295– 300. Jacobsen NJ, Franks EK, Elvidge G, Jones I, McCandless F, O’Donovan MC, Owen MJ, Craddock N. 2001b. Exclusion of the Darier’s disease gene, ATP2A2, as a common susceptibility gene for bipolar disorder. Mol Psychiatry 6:92–97. Jones I, Jacobsen N, Green EK, Elvidge GP, Owen MJ, Craddock N. 2002. Evidence for familial cosegregation of major affective disorder and genetic markers flanking the gene for Darier’s disease. Mol Psychiatry 7:424–427. Karasinska JM, George SR, O’Dowd BF. 2003. Family 1 G protein-coupled receptor function in the CNS. Insights from gene knockout mice. Brain Res Brain Res Rev 41:125–152. Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES. 1996. Parametric and nonparametric linkage analysis: A unified multipoint approach. Am J Hum Genet 58:1347–1363. Lenox RH, Wang L. 2003. Molecular basis of lithium action: Integration of lithium-responsive signaling and gene expression networks. Mol Psychiatry 8:135–144. Morissette J. 1992. L’informatisation de l’information en génétique humaine. In: Knoppers BM, Cadiet L, Laberge CM, editors. La génétique humaine de l’information à l’informatisation. Montréal: Thémis & Litec. pp 89–99. Morissette J, Clépet C, Moisan S, Dubois S, Winstall E, Vermeeren D, Nguyen TD, Polansky JR, Côté G, Anctil J, Amyot M, Plante M, Falardeau P, Raymond V. 1998. Homozygotes carrying an autosomal dominant TIGR mutation do not manifest glaucoma. Nat Genet 19:319– 321. Morissette J, Villeneuve A, Bordeleau L, Rochette D, Laberge C, Gagné B, Laprise C, Bouchard G, Plante M, Gobeil L, Shink E, Weissenbach J, Barden N. 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 88:567–587. O’Connell JR, Weeks DE. 1995. The VITESSE algorithm for rapid exact multilocus linkage analysis via genotype set-recording and fuzzy inheritance. Nat Genet 11:402–408. Owen MJ, Holmans P, McGuffin P. 1997. Association studies in psychiatric genetics. Mol Psychiatry 2:270–273. Roberts SB, MacLean CJ, Neale MC, Eaves LJ, Kendler KS. 1999. Replication of linkage studies of complex traits: An Examination of variation in location estimates. Am J Hum Genet 65:876–884. Sachs AB, Davis RW. 1990. Translation initiation and ribosomal biogenesis: Involvement of a putative rRNA helicase and RPL46. Science 247:1077– 1079. Sham PC, Curtis D. 1995. Monte Carlo tests for associations between disease and alleles at highly polymorphic loci. Ann Hum Genet 59:97–105. Sham PC, Lin MW, Zhao JH, Curtis D. 2000. Power comparison of parametric and nonparametric linkage tests in small pedigrees. Am J Hum Genet 66:1661–1668. Shink E, Morissette J, Barden N. 2003. Genetic heterogeneity in a very large bipolar affective disorder pedigree from Quebec. Am J Med Genet 119B: 65–68. Shink E, Morissette J, Sherrington R, Barden N. 2005. A genome survey supports prior evidence for a susceptibility locus for bipolar disorder on chromosome 12. Mol Psychiatry (in press). Shink E, Harvey M, Tremblay M, Raymond C, Labbé M, Gagné B, Barden N. 2005. Exclusion of non-synonymous snps and a polyglutamine tract in SMRT/N-COR2 as common deleterious mutation for bipolar disorder in the Sagnenay-Lac-St-Jean population. Am J Med Genet (in press). Spitzer RL, Williams JBW, Gibbon M. 1987. Structured clinical interview for DSM-III-R. Biometrics Research Department. New York: State Psychiatric Institute. Stephens M, Smith NJ, Donnelly P. 2001. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 68: 978–989. Suarez BK, Hampe CL, Van Eerdewegh P. 1994. Problems of replicating linkage claims in psychiatry. In: Gershon ES, Cloninger CR, editors. Genetic Approaches to Mental Disorders. Washington, DC: American Psychiatric Press. pp 23–46. Tabor HK, Risch NJ, Myers RM. 2002. Opinion: Candidate-gene approaches for studying complex genetic traits: Practical considerations. Nat Rev Genet 3:391–397. Toyota T, Watanabe A, Shibuya H, Nankai M, Hattori E, Yamada K, Kurumaji A, Karkera JD, Detera-Wadleigh SD, Yoshikawa T. 2000. Association study on the DUSP6 gene, an affective disorder candidate gene on 12q23, performed by using fluorescence resonance energy transfer-based melting curve analysis on the LightCycler. Mol Psychiatry 5:489–494. Tsuang MT, Faraone SV. 1990. The genetic of mood disorder. Baltimore: Johns Hopkins University Press. Weeks DE, Lathrop GM. 1995. Polygenic disease: Methods for mapping complex disease traits. Trends Genet 11:513–519. Wise A, Foord SM, Fraser NJ, Barnes AA, Elshourbagy N, Eilert M, Ignar DM, Murdock PR, Steplewski K, Green A, Brown AJ, Dowell SJ, Szekeres PG, Hassall DG, Marshall FH, Wilson S, Pike NB. 2003. Molecular identification of high and low affinity receptors for nicotinic acid. J Biol Chem 278:9869–9874.