Molecular genetics of schizophrenia: a reviewof the recent literatureDouglas F. Levinson

Purpose of review

The recent literature on the molecular genetics of schizophreniais reviewed, to familiarize the reader with several importantdevelopments as well as a broad range of research efforts in arapidly progressing field.Recent f indings

New genome scan projects, seen in the light of previous scans,provide support for schizophrenia candidate regions onchromosomes 1q, 2q, 5q, 6p, 6q, 8p, 10p, 13q, 1 5q and 22q.Linkage disequil ibr ium mapping studies of several of theseregions have produced evidence from relatively large samplessupport ing the associat ion of schizophrenia to neuregulin-1(NRG1, 8p21-p12) , dysb ind in (DfNBPt , 6p22.3) , pro t inedehydrogenase (PRODH2, 22q11.21), G72 (13q34) withweaker evidence implicating its interacting gene o-amino acidoxidase (DAAO, 12q24), and catechol-O-methyltransferase(COMT,22q11.21). Other reports have described including theapplication of microarray techniques to schizophrenia post-mortem tissue, candidate gene studies in diverse regions,efforts to develop quantitative phenotypes (e.g. neuroimagingand neuropsychological variables) and proposed models ofschizophrenia pathogenesis.

Summary

Schizophrenia l inkage f indings are beginning to converge on anumber of chromosomal regions. Linkage disequil ibr iummapping studies are beginning to produce f indings of greatinterest in some of these regions, and addit ional f indings shouldbe expected. Enlarged linkage and association samples,combined with rapidly evolving technologies, hold out thepromise that in the next 5-10 years, the role of some specificschizophrenia susceptibi l i ty genes wil l be confirmed result ing inan initial understanding of the pathogenesis of schizophrenia.

Keywords

schizophrenia, genetics, genetic l inkage, genetic associat ion,candidate genes

Curr Opin Psychlatry 16:157-170. rr 2OO3 Lippincott Williams & Wilkins.

Department of Psychiatry, University of Pennsylvania School of Medicine,Philadelphia, PA 19104-3309, USA

Correspondence to Douglas F. Levinson, MD, Department of Psychiatry, University ofPennsylvania School of Medicine, 3535 Market Street, Room 4006, Phi ladelphia, pA19104-3309, USAT e l : + 1 2 1 5 7 4 6 5 1 5 1 ; f a x : + 1 2 1 5 7 4 6 5 1 5 5 ; e - m a i l : d f l @ m a i l . m e d , u p e n n . e d u

Current Opinion in Psychiatry 2003, 16:157-170

Abbreviations

ASP affected sibling pairLD l inkage disequil ibr iumNIMH National Institute of Mental HealthNMDA N-methyl-o-asparlateSNP single-nucleotide polymorphismUTR untranslated regionVCFS velo-cardio facial svndrome

:a., 2003 Lippincott Wil l iams & Wilkins0951 -7367

lntroductionThis art icle wil l revicw new l i terature relevant ro rhemolecular genetics of schizophrenia from approximatelythe middle of 2001. Remarkable progress has been madein the 15 years since serious investigation began in thisf ield. The past year alone witnessed the publication ofeight new genome scans [,"-6",7' ,8"] and a meta-analysis of publ ished genome scans [9"], rvirh evidencethat gcnome scan data are starting to converge on a sct ofchromosomal regions; f ive reporus of substantial evidencefor the associat ion of schizophrenia with specif ic genes inposit ional candidate regions [10"-14"], and rwo reporrsof thc use of microarray technology to screen for genesshowing dif ferential expression in the brains of schizo-phren ia pat ients [15 ' ,16" ] , one of which detectcd a genefor which evidence for generic associat ion was alsoobserved l l7"l . I t appears l ikely that posit ional cloning,microarray and other tcchnologies wil l soon produccrcpl icable f indings that wi l l begin ro elucidate thepathophysiology of this severe comn-ron disease.

Progress in f inding genes through posit ionalcloning methodsPositional cloning is the primary srraregy available forfinding susceptibil i ty gcnes for disorders with no knor,vnpathoprhysiology. The strategy is summarized in Table 1.

Multiply affected families are screencd with a DNAmarker map of all chrornosomes (genome scan). Statisticalanalyscs consider whether i l l relativcs have inherited thesame marker alleles, indicating the proximity of a diseasesusceptibil i ty gene. For complex disorders (such asschizophrenia) that probably have multiple interactingsusceptibil i ty gencs, genome scans can only locate gcneswithin 10 000 000-30 000 000 base pair (bp) regions.

In thcsc candidatc regions, addi t ional markers aregcnotyped 1-? cM apart, which somerinlcs narrou,s thecandidatc region.

DOI: 1 0. 1 097/01 .vco.000005861 7.61 505.a5 157

158 Schizophrenia

Table 1. Stages of positional cloning studies

Stage Cl in ical samole Molecular strategy Outcome

Genome scan

Fine-mapping

LD mapping

Functional genomics

Multiply affected pedigrees

Multiply affected pedigrees

Cases versus controls orcase-parent trios

Animal models (e.9. 'knockout '

mice), cases, postmorlem tissue

Genotype microsatell ite DNA markersa t 5000000-10000000 bo(5-10 cM) spacing, or denserSNP markersGenotype microsatell ite markers incandidate regions aI 1-2 cM spacing(or denser SNPs)Genotype markers (primarily SNPs) at10000-50000 bp (10-50 kb) spacing

Physiological studies to establish therole of the gene and protein, effects ofmutations on physiology and behavior,and response to treatments for thedisease

Evidence for l inkage in 1O-30 cMcandidate region(s)

Evidence for l inkage in 2-10 cMcandidate region(s)

Evidence for association of small setsof adjacent markers with disease,implicating one or two specific genesPlausible evidence that the gene playsa role in disease susceptibil ity;detection of other interacting genesand oroteins

LD, L inkage disequi l ibr ium; SNP, s ingle nucleot ide polymorphism.

Linkagc d isec lu i l ib r ium (LD) mapping is bascd on adiffercnt principlc: mosr of the gcnome's DNAscqucnce variat ions have arisen only oncc or a fewt imes. Somc 'unrc la ted ' i l l ind iv idua ls may havcinherited from a single ancient ancesror an identicalDNA scgment that incrcascs rhe r isk of diseasc. Eachpaticnt 's complcte DNA sequencc cannot yer bescquenced cost-effect ively, but one can study manyvariat ions (single-nuclcotide polymorphisms; SNPs)10000-50000 bp (10-50 kb) apar t in the candidatcrcgion in 'unrelate

d' cases ro look for 'haplotypes'

(adjaccnt variat ions on rhe samc chromosome) that aremore common in i l l ind iv idua ls . I t is be l ieved thatsomc (not al l) susceptibi l i ty variants underlying com-non complex disorders can be identi f ied in this waywith larger samples, improved SNP maps and cvolvingmolecu lar mcthods.

Functional studies can rhen bc init iated to determincthc genc's function, inrcractions and possiblc rclat ion-ship to t l-rc disease.

Gencral ly, two gcnome scan strategics have becnemployed. One is to study srnal l samples of famil icswith as many i l l relat ives as tr-rossiblc, in the hopc thatone or a few genes are conferr ing rnost of the r isk ofd iseasc [18" ,19" ] . The other is to co l lec t many farn i l ieswith at least an affccted sibl ing pair (ASP), assuming thatmult iple intcrzrct ing gcnes each confer a small proport ionof r isk [20], so that hundreds to thousands of famil ies rvi l lbc nccded t?t1. Thc complexit ics of this debate arebeyond thc scopc of this paper. Exist ing data abotrtfan-r i l ial pattcrns of schizophrenia tend to slrpporr rhesecond vicw [20]. Srnal l samples tend to exaggcrarc rhegcnetic cff-ects of somc loci whilc missing orhcrs entirely

[22,23]. Howevcr, somc of chc bettcr-su1-rportcd l inkzrgefindings in schizophrenia wcre init ial ly detectccl in srnal lsamples 124,251.

New genome scans

Eight new scans and one part ial scan \ l /erc publ ishedthis past ycar. Regarding the interpretat ion of scanresults, 'gcnome-wide

signif icance' usually refcrs to aresult that is expected to occur by chancc once in 20scans, and 'suggcstive

signif icance' refers to a resultthat would occur once per scan on average t261.However, for complex disorders, stat ist ical ly signif icantl inkage is not rel iably obscrved at feasiblc samplesizes, and the pattern of results across studies mayprove to bc more important. In this discussion, 'narrow'

diagnoses rcfcr general ly to schizophrenia and schizo-affect ive disorder, and 'broad' or 'spectrum' diagnosesrefer to schizophrenia-related psvchoscs and personali tydisorders.

Straub et o/. [1"] publ ished scan results for 270 kishpedigrees. They prcviously reported evidence forl inkage on chromosomc 6p2l-24, which approachedgcnome-wide signif icance dcpending on horv onecorrccts for mult iplc tests [27], and suggesrive evidenceon 5q21-31 [28], Bp22-21 t29] and 10p15-p11 t301.Results suggested at least two suscepribiliry loci onchromosome 6p, both of which have been supported byother studies (see below), but peaks this closc rogerherare diff icult to provc with currcnr methods. Thisrelatively large samplc was drawn from one ethnicpopulation, which might improve the power to dctectl inkage [31]. But the scan was completcd some years agousing threc different, sparse (20-30 cN,,I) maps insubsamples of 90 farnil ics. A range of diagnostic andgenetic models werc analysed to cxplore rhe data fullv,btrt this also makes thc findines more diff icult rointcrprct.

Delisi et a/. l2"l published rwo scans. The largestschizophrenia scan to date was of 291 fan-ri l ies (333indcpcnclcnt ASPs). Tu,o slrggesrive findings \\rcre

obscrved, on chromosornes 10p and Z (centromcricregion), lr, ith more modest evidence on 22q, The sccondscan was of 99 families from the Ccntral Vallcy region ofCosta Rica (62 ASPs by a narrow diagnostic n-rod el, l0Zby a spectrl lm modcl) [3"]. The maximum evidcnce forl inkagc (fall ing short of the suggesrivc level, as has beenthc casc for a number of scans) \\ras obsen,cd inchromosome 5q34.

Paunio et a/. f4"] analysed scan data for two subsarnples;163 Finnish gencral populat ion fami l ies (191 ASPs), and47 nuclear families (60 ASPs) from an isolated and inbredrcgion. This group had previously reported suggesrivecvidence for l inkage in two differcnt regions of distalchromosome 1q, one in the national and the other in theisolatc sample t321. Thc scan prodr-rced significantf indings on distal 2q ( isolate sample) and on 5q31(national sample).

Gtrrl ing et a/. [5"] studied 13 Brit ish and Icelandicpedigrees. Suggestive l inkage was observcd on chromo-somes 1q32.2, 5q33.2, BpZl-Z?, l lqZ3.2-q24 and?0q12. t -11 .23 .

Camp et a/. 16"] studied scvcn extendcd pedigrecs fromthc isolatcd island nation of Palau. Significant l inkagewas obscrved on chromosomcs 2pl3-14 (as reportedpreviously I33l) and 13q12*22, and suggcsrive l inkageon SqZZ-cryer.

Garver et a/. 17'l studied 30 US pcdigrees. No suggesriveresults were observed. The most positive scores were onchromosomes 1p, 5p, 20q and Bp.

Lindholn'r et n/. [8"] studicd a set of pedigrees fromnorthern Sweden (39 narrow and four broad cases) thatwcre intcr-relatcd going back 1Z generations. Significantl inkage was observcd on chromosome 6q25. A six-marker haplotype segregated was observcd in most i l lrelatives. Additional peaks includcd chromosomes 3p25,20p11.2, 6p24 and 8p21-22.

Maziade et a/. [34"] reported a parrial scan (denser incandidate regions) in 19 Quebecoise pedigrees withschizopl"rrenia or bipolar disordcr. Suggcstive l inkage toschizophrenia was observed on chromosomcs 6pZ2-24and 18q.

Although most scans to nor replicatc each other, therc isin fact substantial convergence in the rcsults across allstudies. Prcvious l inkage findings have bccn reviewedreccntly by Watcrwort et a/. f l8"] and Baron [19"], andr,r. i l l not be thoroughly rcvicwed hcre, but there issupport f ion-r sevcral studics for (at least) regions ofclrromosomcs 1c1, 2c1, 5q, 6p, 6q, 8p, 10p, 13c1, 15c1 zrnd22q. t \ mcta-analysis of ptrbl ished schizophrcnia l inkage

Molecular genetics of schizophrenia Levinson 159

reports (including some but not all of the above data)reported P values < 0.001 in candidate regions ofclrrcrmosomes 1q, Zcl, 8p, 13q and 22q 19"1. (A largermeta-zrnalysis is being carried out by a collaboration ofinvestigators and should be availablc soon.) Pou'eranalyses suggest that replicable l inkagc u'ould requiremultiple samplcs of 500-1000 ASPs or more. Tu,osamples of 500 and 900 ASPs are clrrrenrly beingcollected for thc National Institute of Mcntal Health(NIMH) ccll repository. A series of multicenrer colla-borations in which samples of 500-800 pedigrees weregenotypcd produced evidence supporuing l inkagc onchromosomes 6p [35], 6q [36], 10p t36l and Bp [35],al though not 1q 137'1,5q [36] or 13q [36]. I t is l ikely thatthat some or perhaps many of these candidate regionscontain schizophrenia susceptibil i ty loci, cach contribut-ing to a small increasc in population-wide risk (althoughsome loci could have larger effe cts in indir. idualfami l ics) .

Candidate genes in positional candidate regionsFor the first t ime, LD mapping studies of positionalczrndidate regions have implicated specific ge nes inschizophrenia susceptibil i ty. Previously, no strong andreplicable candidate genes had emerged from studies ofgenes involved in antipsychotic drug mechanisms (c.g.dopamine receptors) or in central nervous svsremprocesses that could be hypothetically related to schizo-phrenia. In the absence of evidence for l inkage in aregion, thc prior probabil ity for the association of a gencis so low that statistical proof becomes exrremely diff icultt381. These new reports (Table 2) are thereforeimportant new developmcnts. F'urther replication wil lbe needed before the validity of these associations canbe evaluated.

NRGlA genome scan in 33 Icelandic pedigrees with 105cases of schizophrenia produced suggesrive evidencefor l inkage on chromosome 8p [10"], near wheresuggestive l inkage has been reported by others[18",19"] . LD mapping studies in cases and controlssuggested thc association of individual SNPs and SNPhaplotypes in MGI (neuregul in- l , 8p21-pl2). The'core' haplotypc was observed in 15/% of 478 affectedcascs and 7.5% of 394 controls, wi th P:0.000087 forcascs not known to be related. IVRGI is a glycoproreinwith a variety of isoforms that bind to the ErbB farnilvof tyrosine kinase transmembrane receprors. NRGI andErbB4 l-reterozygous mutant mice demonstrated hvper-activity that was rcduced by very lor,v doses ofclozapine, slightly dcficient pre-pulse inhibit ion, anda small reduction in functional I/-rnetl-rvl-D-aspartate(NN,IDA) reccptors. Note that IVRGI lies 9-25 cN,Itowards the centromere from previously rcported 81-,l inkagc pcaks, and i t is possible rhar anothcr gene in

160 Schizophrenia

Table 2. schizophrenia candidate genes in positional candidate regions

Gene(s) Location Study population(s) Comments

Neuregul in 1 (NRGl) [10. . ]

Dysbindin (DTNBPT) [1 1. . ]

Proline dehydrogenase(PRODH2)

G72 (o-amino acid oxidase;DAAO) [13.{

Catechol- O-methyltransferase(coMl) 114--l

8p12-21

6p22

22q11

13q34 (12q2a)

22q11

lcelandic. Linkage: 33 pedigrees.Association: 478 cases versus394 controls

lrish. Linkage: 27O pedigrees.Association: same (family-basedmethod)

107 US tr iads (adul t schizophrenia) ;29 US childhood schizophrenia triads;75 Afrikaner cases and .'|09

controls

213 French Canadian cases and241 contro ls; 183 Russian cases and183 contro ls

Ashkenazi. Association : 7 1 4-7 24 casesand 2849-4899 controls

NRGl lies outside the 8p region wherepositive linkage evidence has beenreported, and might not explain thoserepons.The gene product is part of the dystro-phin protein complex involved in muscu-lar dystrophy, but may also be involved insignal transduction andreceptor gene expression.Association was observed in early-onsetcases. Sample sizes were small and Pvalues were modest (= 0.001), butpositive evidence was observed in threesamoles.G72 was identified through LD mappingof l inkage regions from other reports.Yeast two-hybrid experiments identifiedDAAO as interacting with G72; modestassociation was observed for DAAO.COMT is in the VCFS deletion region.It degrades catecholamines includingdooamine.

LD, L inkage disequi l ibr ium; VCFS, velo-cardio facia l svndrome.

t l rat rcgion, rarher than a variat ion in IVRG/, wi l l provcto cxplain thosc f indings.

DINBPl

Straub el o/. [11"] reported the associat ion of schizo-phrcnia to SNPs and SNP haplotypes on chromosome6 ; t22 .3 , imp l i ca t i ng t hc DTMPI (dysb ind in ) gcne , i nthe l inkage region from their and orhers' genome scans[18" ,19" ,? ,71. Dysbind in b inds to dyst robrev in , par t o fthe protcin complex involved in the pathogenesis ofmuscular dystrophy. Thesc proteins have diverse func-t ions rclated ro ncurotransmitter signal transduction.Evidence suppoming this associat ion has recently beenrcported in two German samplcs [39].

PRODH2

Individuals rvith velo-cardio ftrcial syndrome (VCFS) areat a strbstantial ly increascd r isk of schizophrenia l401Zl,so that the idcnti f icat ion of the relevant gcnes in thc221111 microdelct ion region is an important task. Liu r/a/. l lz" l rcportcd the associat ion of schizophrenia toSNPs in PRODHZ (prol ine dehydrogenase, ZZctl l .Zl),part icularly in chi ldhood cases and in aclults with agc atonsct bclow thc age of 18 yeirrs. Eviclence for associat ionwas modest, but was obscrved in three scparate smallsan'rplcs. Prol ine dehydrogcnase is a mitochrondrialcnzymc involvcd in transferr ing reclox porential acrossthe mi tochondr ia l membranc.

G72 and DAAO

Chrrmakov et n / . [13" ] in i t ia ted LD rnapping s tuc i ies o fchromosome 13q34 in thc region lvherc signif icant

l inkage was reported in rwo samples [25,431. Strongassociat ion was observed in one French Canadian samplcbetween schizophrenia and SNPs in a novel genc, G7Z,with weaker associarion in a Russian samplc. Yeast two-hybrid expcriments and subscquent studies demon-strated that G72 acrivares D-amino acid oxidase (DAAO,12q24), an enzyme that oxidizcs D-serine, an acrivaror ofNN4DA receptors. Weak associat ion was observcdbetween schizophrenia and DAAO.

COMT

Catechol -O-methy l t ransferase (COMT, 22q|LZ1, in theVCFS region) is one of the major degradative parh-ways for catecholamincs including dopamine. Associa-t ion with schizophrenia has been inconsistent for aVal/Met polymorphism [18"], with several recenraddit ional ncgativc reporrs 144'46',47,48.1, althoughtwo of the studies reported modest associat ions forsymptom severity 144'l or age ar onser 146.1.Recently, Shifman et a/. [14"] screcned new COIITSNPs with pooled genotyping in morc than 700Ashkenazi Jewish cases and 3000-5000 controls, rvi thindividual genotyping confirming srrong associat ionsfor trvo SNPs (P:0.00016-0.00003). Weinberger et o/.

[49'] proposed that CO,+IT scqlrence variation modificscognit ion through cffects on dopaminergic transmissionin prefrontal cortex.

Such reports usher in a new era in schizophre niagene t ics. One might cxpecr more such repofts roLlppear ovcr the next 3-5 years, u, i th increasingly largesamplcs support ing thc associat ion of schizophreni ir

rvith plausible candidate gcnes in l inkage candidateregions, or gencs idcntif ied by large-scale exprcssionstudies. Init ial rcports may bc diff icult ro interprct:many tests musr be performcd on SNPs and haplo-types, and there is no definit ive threshold for ^significant f inding. Data rnining tcchniques arc evol-ving in this field, and rcplication is l ikely to be thecorncrstone of data intcrpretation. Thc allthor u,ouldprcdict that, despite these diff iculties, the pattcrn ofrcplication studies and physiological f indings wil lult imately become clear and convincing for ar lcastsome of thc new candidate gcnes. Throtrgh thisprocess, the first real clues ro rhe parhophysiology ofschizophrenia are l ikely ro cmcrgc.

Other studies of candidate regions and genesOthcr notcworthy reports include applications of micro-array technology and studies of candidate gencs inph armacological ly-relevant ncurotrans rni tter system s andin posi t ional candidate regions.

Probing the genome with microarraysSeveral sttrdies demonstrated the emerging porcntial ofmicroarray methods to probe the genome or protcomefor candidate pathways. The decrcased expression ofRGSI (regulator of G-protein signall ing 4, lqZl-q?D hadbeen observed in a complemenrary DNA microarraystudy of brains from schizophrenia and control indivi-duals [15']. SNPs were idcntif ied in the region, zrndhaplotype analysis revcalcd four SNPs with a modestassociation to schizophrcnia in two indcpendenr sam-ples, wi th a t rend in a rhird sample 117").The gene islocated in a l inkage candidate region 15",241. In asecond study, Hemby et a/. [16"] probed 1B 000messengcr RNAs in case and control brains anddemonstrated significant differences in the expressionfor a numbcr of genes. A further example is discussed forchromoso me 22q (below).

Genes related to dopaminergic neurotransmissionSchizophrcnia has nor been convincingly associatedwith polymorphisms in genes rclated to doptrminergicfunction, although meta-analyses have suggested a smallbut significant association for homozygosity at a ;roly-morplrism in DRD3 (3q13.3) [50]. Therc are rwo rcccnrreports of slightly positive evidencc for an association atDRDZ (I1c122-q23) ln l ' ,52'1, and one repon [53' ] wasnegative. Therc was a rcport of thc association(P = 0.0001-0.05) betwccn measures of eye-trackingdysftrnction in schizophrenia patients and DRD3 [54'],as well as a negative study of DRD-I homozygosity andof several new polymorphisms in alternarivc promoterand 5'untranslated regions ( t iT l {s) in the gene [55' ] . Afurthcr report fonnd a moclest association bcnvecnschizophrenia and two markers in or near DIIDS (ap16)

[5(r ' ] .

Molecular genetics of schizophrenia Levinson 161

Genes related to serotonergic neurotransmissionA rnodest but significant association bcru'een schizo-phrcnia and a polymorphism in the serotonin-2Areceptor gene (HTRZA, l3ql4-q21) rvas reported in alarge multicenter analysis t57) and a mera-analvsis[18",19",58]. Therc are two new negar ive reports[59,60]. Association had been reported in Han Chincseindividuals for the serotonin transporucr gene (SLC6A,formerly 1HTT, 17qll. l-q12) [61], and a very modestreplication (P :0.043) was rcported in the san-repopulation 162'1. Negative studies of 5HTRI B (6c113)

t63l and 1HTR3A (l lq23.l-c123.2) 164l have beenpubl ished.

Genes related to glutamatergic neurotransmissionA numbcr of l ines of cvidence have suggested a role forglutamatergic dysfunction in the pathogenesis or phar-macology of schizophrenia [65']. A modest association(P:0.0105) was rcported betwcen schizophrenia andGRIK3 (ionotropic glutamate recepror kainate 3, 1p34-33) in 99 cases and 116 controls [66' ] ; and an associat ion(P:0.0022) was reported for GR,M,I (metabatropicglutamate receptor 3, 7q7l . l -qZl .Z) in 265 cases andZZ7 controls, which could not be replicated in anaddi t ional 2BB cases or 128 tr ios [67' l .Negat iveassociation results were reportcd for GR'M4 (meta-botropic glutamate receptor 4, 6p21.?) [68'], GRIKI( ionotropic glutamate rcceptor kainate 1, 21q22.11)[69'] and GkM2 (metabotropic glutamate recepror Z,3ptZ-p11) [70'].

Chromosome 1qSignificant l inkage was rcported on chromosome 1q21-q22 (ry 158 MB from pter on chromosome 1) in a 'Cel t ic '

Canadian sample lZ+1, and suggestive l inkage wasreported on lc lZl-qZ3 (1165 MB) in Icelandic andBri t ish pedigrees [5"] , on lq41-q42 (x208 MB) infarnil ies from a generic isolate in Finland f5",321, and on1q42.1 (x230 MB) in a Finnish general populat ionaffected sib-pair sample [5",321. The latter regioncontains the breakpoint of a balanced (1;11) (q42.I :q14.3) translocation l inked to psychotic and mooddisorders in an extended Scottish pedigree, rvithmaximum lod scores of 3.6 for schizophrenia, 4.5 formood disorders and 7.1 for both [71]. The DISCI andDISCZ genes are disrupted by this breakpoint [72"], butno variation associated with schizophrenia has ver bccnfound in these gcnes 173'1. The genes in thc l lqbreakpoint region have also been identif ied [74'].Linkage was not supported in a multicenter samplc clf779 pedigrees using a 5 cNt[ markcr map of lq2l-q42[37'1, or in 1q21-(PZ in extended French Canadianpedigrces [75']. There have been mixed reporus regard-ing thc association with thc CAG repear polvrnorpl'r isn-rin KCI{AB ( lqZI.Z) [18"] . A new Japanese study u 'asncgative [76'l; but in Israelis, longer alleles wcrc

162 Schizophren ia

itssociated with morc severe negarive and parzrnoidsymptonts [77 ' ] .

Chromosome 6N,{oclest irssociation was reportcd for SNPs in TIYFZA(tumor nccros is fac tor i r lpha, 6p21.1-21.3) [78 ' ] . A h igh lysignif iczrnt associzrt ion had bccn reported in NOTCH4(6p21.3) in a srna l l sample 1791, but there havc beenmany non-rep l icar ions [80,81,82 ' ,83 ' ,84] . A suggcst ivel inkage was reportcd in t l-re acl jaccnr HLA rcgion(6p2l.Z) [85'] . I- Iowever, although an imn"runologicalro le in the pathogenes is o f sch izophren ia is p laus ib le

[18" ,19" ] , no assoc i : r t ion rvas observed in recent repor tson c lass I a l le lcs in Japanese cases and cont ro ls [86 ' ] , inclass II DRll l , DOAI, DOII| and DPBI al lclcs in HanCl-r inesc tr ios [87'] , or in DOB 1, DRB t, DgAl, or HLA-Aal le les or haplorypcs in pr imar i ly European s ib-pa i rfanri l ies and tr ios [85'] . Reports on DTNBPI and on (rc1

l inkage have bccn discussed above.

Chromosome 8pNo associat ion with schizophrcnia was obscrved formarkers in three plausible candidate gencs (PIVOC,

CHRAIAZ and lVAZ1) [U8'] in the 8p2l-22linkage rcgiondiscussed above.

Chromosome 15qFrcedman and col lcaglres [89] dcmonstrated the highlysignif icant l inkage of chromosomc 15q13-q14 to defi-cicnt inhibit ion of thc P50 auditory evoked porential(pre-pulse inhibit ion) in schizophrenia probands andthcir affccted and unaffected rclat ives, and showed thatt lrc alpha-7 nicotinic receptor gcne (CHRArAT l wasinvolved in this defici t in animals and was a plausiblecandidate genc for schizophrenia t90-921. Howcver,evidence for l inkage ro schizophrenia in this regionhad becn modest. Now, a re-analysis of the NIMHSchizophrcnia Gcnetics Init iat ive genome scan using adominant paramctric lod scorc analysis produced sig-n i f icant ev idcnce for l inkage on 15q13-q14 [93" ] .Suggestive evidence for l inkage was also reportcd nearCHRMZ in

' I 'a iwancsc fami l ies f94 '1 , w i th more modest

support in two Amcrican samplcs [95',96'] . An associa-t ion was also observecl (P: 0.0004) bcnveen mtrrkcrs arCHIIATAT and schizophrcnia in 31 pedigrees frorn theAzorcs 197'1, but not in pedigrccs with thc periodiccatatonia phenotype, which had shown l inkage to thisrcg ion [98 ' ] . On thc bas is o f thcsc resu l rs , 15q13-q14cmergcs as a strongcr schizotr-rhrcnia candidatc rcgion.

Chromosome 22q

Microdclct ions of 22c111.21 (VCFS) rcpresenr rhc onlycytogcnctic abnormali ty that is clcarly associatccl rvi th asubstantial r isk of chronic schizophrenia, suggestir-rg thartonc or more gcnes in this rcgion could play a role insch izophrcn ia in the abscncc of dc lc t ions. A l though

cvidcnce fbr l inkage h:rs been modest, a mcta-analysissuggcsted significant l inkagc [9"]. Evidencc forPRODHZ and COMT \vas discussed above. VCFSclclctions were again found to be rare in schizophrenia- one in 300 Japanese cases [99' ] . For genes in thedclction region, moclest evidence for association \\/asobservcd for two polymorphisms in promoter regions,including UFDIL (ubiquitin fusion dcgradation 1-likc)in two small sarnples [100'], and SNAP29 (synaptosomal-associated protc in, 29000 M,) [101' ] ; and for SNPs inZA|F74 (zinc finger protein 74) ro agc-at-onser bur norschizophrcnia in 300 Japancsc cases and 300 controlsll0z'1. The latter agc-rlt-onset f inding was replicated inanother 769 cascs |02' | Chi ldren with VCFS hadsin-ri lar bchavioral and psychiatric disorders to childrcnmatchcd for the level of cognitive function, lcading theauthors to doubt the specific rclationship betwccn VCFSand schizophrcnia [103' ] , but th is conclusion ignorcs the20Vo prevalencc of schizophrenia svmptoms amongadults u, i th VCFS t411. Outsidc the delet ion region,Ntleyer et a/. [104'] described a purarive porassiumchannef gcne, WKLI (AF319633_1, 22q73.33), andreported that a missense muration in this gene co-segregated with pcriodic catatonia in one extendedpedigree. Three groups failed to find the mutation orassociation to other SNPs in the region in schizophreniasamples [105' ,106' ,107]. In per iodic cararonia, an asso-ciation was not observed for CELSRI (cadherin EGF (7cpidermal growth factor-l ike repeats) LAG (2 lamininAG-type repeats) seven-pass G-type recepror 1,22q1.3.3)[108' ] . Final ly, Mimrnack el o/ . [109' ] carr ied our acDNA microarray-based gcne expression analysis ofschizophrenia and control brains from ru.o collections,and in both they found an increased expression ofAPOL|, APOLZ and APOL4 (apolipoproteins 7,2 and 4,all in 22qlZ outside the VCFS deletion region). Tl 'resefindings suggcst a possible pathway in schizophrcniapathclgenesis.

Other candidate genesA role in schizophrenia has been hypothesized forcholecystokinin (CCK, (3p22-p21,.3), which modulatesdopaminergic neurotransmission [110' ] . Recenr rcporrsinclude slightly positive evidence for the association ofscl-rizophrenia with a polymorphism in the promorerrcgion in US cases [111' ] and u, i th polyrnorphisms in thepromotcr region of the CCK-A recepror genc (CCKAR,4p15.1pf5.2) in Japanese cascs and controls [110' ] , butnot to a microsatcllite in the CCK-II rccepror gcne(CCKIl l l , 11p15.4-p15.1) in Japanese cases and controlsl l l2 l . An associat ion (P:0.031-0.0001) u 'as reportcdbetu'ccn schizophrcnia and threc SNPs in myo-inositolmonophospl'ratasc 2 (IIIPAZ, 18p11.2), in 302 Japaneseschizophrcnia cerses vcrsus 308 controls and 205 affectivedisordcr cases [113' ] . Nt lodest evidcnce for associ i l r ionwas reportcd for a SNP in CTL|I-1 (c_vtotoric T-

lymphocytc anr igen-4, 2q33) in 116 Korean cases and149 controls (P:0.003), as a result of fcwer hctero-zygotes in the cascs [114']. Sirnilarly, 268 lapancse cascsshowed increased homozygosity at l inked SNPs inGRIATZB (rhe Ar,4,lDA-28 recepror, l?plZ) versus 337controls (P:0.004) [115' ] . Nlaes et a/ . [116' ] revieu'edcvidence for alterations in thc acutc phase inflammzrroryresponsc in schizclplrrcnia, and reportcd an association ofpolymorphic var iat ion in haptoglobin (HP, I6q22.1) andhaptoglobin plasma levcls in 98 Italian cases versusestabl ished distr ibut ions in the local populat ion. Final ly,Hong et a/. l1l7'l observed a modcst associationbetu,een schizophrenia and TPH (tryptophan h1'droxy-lasc, 11p14-p15.3) in 196 I{an Chinesc cases and 251controls.

Additional ncgative association studics of candidatcgenes publishcd during the rcview pcriod wil l not bcd iscussed in de ta i l here [118-136] .

Progress in identifying quantitative traits forgenetic studiesTherc werc sevcral reporrs about schizophrcnia-relatcdquantitative traits that showed promisc as phenotypes forgenet ic studies.

Neu ropsycholog ical measu resIn 264 cases and rclativcs from a geographical isolate inFinland, measures of working memory showed srronghcritabil ity, possibly becausc of a small number ofcontributing loci U37'). Measures of memory and IQdeclined before the onser of psychosis in high-riskindiv iduals in thc Edinburgh high-r isk project [138].However, in the same project, attenrional dysfunctiondid not diffe rentiare rhe relatives of schizophrenic versuscontrol subjects [139' ] .

Neuro-imaging measuresIn a crcative analysis from thc Edinburgh project, areduced volume of the amygdalohippocampal complexwas observed in 'obligate carriers' (well individuals witha schizophrcnic sibling and offspring) [140']. In a studyof monozygotic and dizygotic twins discordant forschizophrcnia and control rwins, an Llpward bowing ofthc corpus callosum was also suggested to be a markcr ofvu lncrab i l i t y [141 ' ] .

Neurological s ignsNcurological signs wcrc increascd in thc relatives ofschizophrcnia patienrs, bur the low rclative risk versuscontrols suggestcd rhat thcy might bc poor gcneticmarkcrs [142'). A pattern of dysmorphic featurcs wasfound to diffcrenriare a subgroup of 1B schizophrcniacirscs from othcr clusrcrs in a sample of 123 clinicalschizophrenia cascs and 36 cascs referred for possiblevcFS [143].

Molecular genetics of schizophrenia Levinson 163

Cl in ical d imensionsCardno et a/. [I44'] showed that 'first-rank' symprorns(such as dclusions of bcing controlled or hcaring voicesconversing) u'ere heritable in the Maudsley tlvin series,but lcss so than the category of schizophrenia. Thevalso rcported that generic l iabil i ty ro schizophrenic,manic and depressivc symproms appeared ro ovcrlapsubstantially when rated non-hierarchicallv [145'].Kendler U46'1, in a commentary on rhe latter study,suggested that the data are better interpreted asshowing that manic and depressive syndromes aresomewhat heritablc in individuals rvith schizophrenia,because no genetic overlap was shown betweenindividuals with primary diagnoses of schizophrenicand rnood disorders.

Models for genetic studies of schizophreniaSeveral intr iguing models have been proposed forconccptualizing gcnetic factors in schizophrenia.

Advancing paternal age and de-novo mutationsA re-analysis of data from a birth cohort study demon-strated that advancing paternal age slightly but signifi-cantly increased thc risk of schizophrenia, after adjusringfor possible confounding factors such as maternal ageU47"1. In another study [148' ] , schizophrenia pat ientswithout a known family history of schizophrenia hadsignificantly older fathcrs than those with a familyhistory. Malaspina [I49'] integrated these and othcrdata into an intriguing hypothesis that de-novo germlinemutations can increase the risk of schizophrenia, whichsuggests that novel stratcgies may be needed to identifvthese mutations ll49'1.

Neu rodevelopmental hypothesesl.ewis and Levitt [150"] reviewed the evidence fora neurodevelopmental hypothesis of schizophrenia,suggesting a polygenic-multifactorial model wherebygenctic predisposition alters development, ge ne-environment interactions influence or trigger gencriccffects, and cumulative effects of altered dcvelopmentlead to a relatively stable and treacment-resistanr state.Maynard et a/. 1151'] proposed a two-hit model requiringgenctic or early environmental predisposing factors plustriggcring factors closer to onser. Schiffman et a/. [152"]reported that among 93 high-risk offspring, the numberof physical anomalies at age 11-13 years predictedschizophrenia spectrum disorders at age 31-33 vears.Therc were high rates of anomalies and of spectrumdisorders in the samplcs. Bassett and col leagues [153' ]rcviewed evidcnce for ncurodevelopmental hypothcses,focusing on physical anomalies and the implications ofsyndromes such as VCFS. DeLisi [154'] suggested tl-ratschizophrcnic patients have language clisorders thatcould be related to genes underlying the uniquelvhuman characteristics of spccch.

164 Schizophrenia

ConclusionThe pacc of discovery is accelerat ing in the f icld ofsch izophren ia genet ics . The mosr s t r ik ing dcve lopmentsin thc past year have included the zrpparenr convcrgcnceof new and o lder l inkage data on zr number o fchromosomal rcgions, zrnd the f irst reporrs of LI)rnapping data slrpporring s;recif ic candidatc gcnes inlinkagc candidatc regions. A|RGl, DT'ArBP1, GT2IDAAOand COMZ r.vere all studied in 400 or ntore cases, rnorethan in most previons candidate gene studies. Hou,evcr,inadcquatc powcr remains a prob lem in both l inkagc andLD mapping studies: ideal ly, i t would be prcfcrable tohave scveral l inkagc samprles of 500-1000 pedigrees andsever:r l associat ion samplcs of 1000-3000 cases andcontrols or probancl-parent tr ios. Associart ion data fron-rvery small samples remain dif f icLrl t to interpret.

The next fcw years are l ikcly ro see thc completion ofthe haplotypc strucrure map ( 'HapMap') and a rapiddecl ine in thc cosr of vcry high-throughput SNPgenotyping. Thc HapMap project wi l l identi fy a minimalset of approximately 100 000-500 000 SNPs, whichdefinc local 10-50 000 bp blocks of LD, i .e. the old,corrurron variat ion in thc human gcnome [155-158]. Thehypothesis to be testcd is thac at least somc of thepredisposing DNA sequencc variat ions for commoncomplcx disorders may be in LD lvi th thcse blocks.However, therc may be roo many dif fcrcnt predisposingvariat ions in dif fcrcnt popularions, with l i t t le correlat ionbetwcen thc obscrvable phcnotype and underlyinggcnotypc t1591. I t is l ikcly rhat cost-eff icienr mcrhodswil l be dcvelopcd to detcct al l sequence variat ions incandidate rcg ions [160] , so rhar one might a lso detect ancxcess of rare variat ions in cases. Diverse microarraystrategies offcr addit ional tools to identi fy functional lyrelevant gcnes, proteins and parhways, and althoughthesc wil l not always be causal [159], they are l ikely toyield insights for trcatment and ult imately fcrr bettercausal hypothcses.

An art iclc by a social scientist suggested some conclud-ing ref lect ions: Conrad [161'] cr i t iqued the widesprcadmcdia cndorsement o f 'genet ic

opt imism' , 'wh ich

emphasizcs the inev i tab i l i ty o f d iscover ing genes andthc good outcomcs of those discoveries, whilc negariveor retracted f indings reccivc no attenrion. Scienti f iccol lcagues in genctics and othcr f ields arc prone tosimilar distort ions' (p. 225). However, optimism is farfrom the dominanr rheme in thc scicnti f ic community.Rather, wc sec an unfbrtunarc osci l lat ion: scientists asr,vel l trs the media tcncl to make exaggeratedly posit iveclaims, and then thc 'ncgative

and rctractecl f indings'havc lcd to profound skepticisrn. What is necdccl is ab:r lanccd ancl stcady approach for the long haul. In thisauthor's vierv, the corncrstonc of progress shotr lcl be thclogic of posit ional cloning: b_v ut i l iz ing data frorn l inkage

studics (and perhaps in thc future, microarray studies) in

a cri t ical \ \ 'Lry - cognizant of me thodological shortcorningsand real ist ic about the strcngth of results and the po\\rer

of samples i t should bc possible to move fromcancl idate regions to candidate genes to thc beginnings

of an undcrstanding of schizophrcnia pathophysiolog_y inthe ncxt decade.

References and recommended readingPapers o{ part icular interest, publ ished within the annual period of review, navebeen highl ighted as:' of special interest" of outstanding interest

1 Straub RE, Maclean CJ, Ma Y, et a/. Genome-wide scans of three" independent sets of 90 lr ish mult iplex schizophrenia famil ies and fol low-up of

selected regions in al l famil ies provides evidence for mult iple susceptibi l i tygenes. Mol Psychiatry 2OO2; 7 :542-559.

Genome scan results are reported for 27O lr ish pedigrees. The scan was init iatedand mostly completed when genotyping this large sample was considerably moreexpensive. l t was thus divided into three subsets, each of which was scannedusing a different 2O-30 cM map, and positive regions as well as candidate regionsreported by others were typed in al l subsets. Mult iple diagnostic and transmissionmodels were tested. Results on chromosome 6p21-24 approached genome-widesignif icance, with a suggestion oJ at least two loci in the region, and suggestiveevidence for l inkage was observed on 5q21-31 , 8p22-21 and 10p15-p11.

2 Delisi LE, Shaw SH, Crow TJ, et al. A genome-wide scan for l inkage to" chromosomal regions in 382 sibl ing pairs with schizophrenia or schizoaffec-

t i ve d isorder . Am J Psych ia I ry 2QO2;159:803-812.This is the largest genome scan published to date, including 294 {amil ies ofpredominantly European ancestry with 333 independent ASPs. Suggestiveevidence for l inkage was observed on chromosome 10p15-p13 and around thecentromere of chromsome 2, with a smaller peak on chromosome 22o12.

3 Delisi LE, Mesen A, Rodriguez C, et al. Genome-wide scan for l inkage to" schizophrenia in a Spanish-origin cohort from Costa Rica. Am J Med Genet

2OO2; 114:497-508.Genome scan data are reported for 99 families from the Central Valley region ofCosta Rica (62-102 ASPs depending on the model). Modest evidence for l inkagewas observed in chromosome 5q34.

4 Paunio T, Ekelund J, Vari lo T, et al. Genome-wide scan in a nationwide study" sample of schizophrenia famil ies in Finland reveals susceptibi l i ty loci on

chromosomes 2q and 5q. Hum Mol Genet 2001; 1O:3037-3048.Two genome scans are reported, one of 163 famil ies (191 ASPs) from the generalFinnish populat ion, and one of 47 nuclear famil ies (approximately 60 ASPs) froman isolated, inbred region. Results for chromosome 1q are reported in Ref. [32].Signif icant f indings were observed on very distal 2q ( isolate sample) and 5q31(national sample).

5 Gurl ing HM, Kalsi G, Brynjolfson J, et a/. Genomewide genetic l inkage" analysis confirms the presence of susceptibility loci for schizophrenta, on

chromosomes 1q32.2, 5q33.2, and 8p21-22 and provides support forl inkage to schizophrenia, on chromosomes 11q23.3-24 and 20ql2.1-11.23.Am J Hum Genet 2001; 68 :661-673.

Agenome scan is reporled for 13 Brit ish and lcelandic pedigrees (56 cases withschizophrenia spectrum psychoses and 12 with broader diagnoses). Suggestiveevidence for l inkage was observed on chromosomes 1q32.2, 5q33.2, 8p21-22,1 1 q23.2-q24 and 20q l 2 .1 -1 1 .23 .

6 Camp NJ, Neuhausen SL, Tiobech J, et al. Genomewide mult ipoint l inkage" analysis of seven extended Palauan pedigrees with schizophrenia, by a

Markov-cha in Monte Car lo method. Am J Hum Genet 2001;69 :1278-1289.This is a repod on a genome scan of seven pedigrees from Palau (an old butgenetically isolated population with an elevated - 2o/o lifetime prevalence - ofschizophrenia), with 40 narrowly defined and 45 'spectrum' cases avai lable. A10 cM map was genotyped and l inkage analyses performed using Markov chainMonte Carlo methods to reconstruct haplotypes followed by Non-ParametricLinkage (NPL) and paramelr ic lod score analyses.

7 Garver DL, Holcomb J, Mapua FM, et al. Schizophrenia spectrum disorders:' an autosomal-wide scan in mult lplex pedigrees. Schizophr Res 2001;

52 :1 45-1 60 .A genome scan was per{ormed on 30 US European and Afr ican-Americanpedigrees with 62 schizophrenia cases, 35 other psychoses and 23 spectrumpersonali ty disorders, using a 10 cM map and NPL analysis.

8 Lindholm E, Ekholm B, Shaw S, et a/. A schizophrenia-susceptibi l i ty locus at" 6q25, in one of the world's largest reported pedigrees. Am J Hum Genet

2 0 0 1 ; 6 9 : 9 6 - 1 0 5 .This is one of the most signif icant f indings in schizophrenia l inkage research, in6q25.2 in a large mult igenerational pedigree (actual ly an inter-related set of suchpedigrees) from northern sweden. The relat ionships within this family go back 12generations, presenting a chal lenge for l inkage analysis, and analyses withalternative allele frequencies produced somewhat different P values. This is one ofthe few examples of highly significant results from the study of this type of rarepedigree. The l inkage region is sl ightly distal to that reported by other groups.

9 Badner JA, Gershon ES. Meta-analysis of whole-genome l inkage scans of" bipolar disorder and schizophrenia. Mol Psychiatry 2OO2;7:4e5-411.A meta-analysis was carr ied out on published schizophrenia and bipolar l inkagedata, using a novel method for combining P values for broad regions of l inkage.For schizophrenia, P values of less than 0.001 across studies were observed onchromosomes 1q, 2q, 8p, 13q and 22q.

10 Stefansson H, Sigurdsson E, Steinthorsdott ir V, et at. Neuregulin 1 and" susceptibi l i ty to schizophrenia. Am J Hum Genet 2OO2;21.B7Z-8g2.ln 33 lcelandic famil ies with 105 avai lable individuals with schizophrenia,suggestive evidence for linkage was observed on 8pi2-21 . fhe region wasdensely mapped in 478 schizophrenia cases and 394 controls, and evidence forassociat ion was observed within NRGI (neuregulin 1) (P=0.000062 for a corehaplotype in one of many tests). The f inding is 8-10 cM centromeric to numerousother reports of 8p linkage, raising a question about whether this gene can explainthat previous evidence. Hemizygous knockout mice were produced and exhibitedmodest evidence of deficient pre-pulse inhibition of startle and for a reduction inNMDA receptors. This thus becomes an important candidate gene forschizophrenia.

11 Straub RE, Jiang Y, Maclean CJ, et al. Genetic variat ion in the 6p22.3 gene" DTNBP1, the human ortholog of the mouse dysbindin gene, is associated

with schizophrenia. [Erratum appears in Am J Hum Genet 2OO2;72:lOOZl.Am J Hum Genet 2002; 71:337-348.

ln a candidate region (6p22) supported by several studies in diverse populations,these authors describe evidence for the association of schizophrenia withDINBP1, dystrobrevin binding protein 1 or dysbindin, based on family-basedassociat ion analyses of SNPs genotyped in a mult iplex lr ish sample in which 6plinkage was first described. Although dysbindin is part of the protein comprexrelated to muscular dystrophy, these authors summarize evidence that it may play arole in signal transduction including in the NMDA and i ' -aminobutyric acid systems.This locus wil l receive intensive study in schizophrenia samples in the comingyears.

12 Liu H, Heath SC, Sobin C, et al. Genetic variat ion at the 22q l 1 pRODH2l" DGCR6 locus presents an unusual pattern and increases susceptibility to

schizophrenia. Proc Natl Acad Sci U S A 2002; 99:3717-3722.This group has been studying possible 22q candidate genes for some t ime. Herethey present their most compelling evidence to date: several SNPs in the PRODH2(proline dehydrogenase) gene, in the VCFS deletion region, show association withschizophrenia in three samples: adult schizophrenia patients, chi ldhood schizo-phrenic patients, and an adult replication sample from South Africa (Afrikaners).The association was observed for individuals with early age at onset. P valueswere modest (P= O.OO1 for the strongest associations). The estimated relative riskfor the most strongly associated haplotypes were in the range of 3-5. Theevidence here is not definitive but this is an important hypothesis that will becareful ly studied in the f ield.

13 Chumakov l , Blumenfeld M, Guerassimenko O, et al. Genetic and" physiological data implicating the new human gene G72 and the gene for

d-amino acid oxidase in schizophrenia. Proc Natl Acad Sci U S A 2OO2;99:1 3675-1 3680.

LD mapping studies on chromosome 13q34 were init iated because of previousrepods of significant linkage. SNP haplotypes in a novel gene, GZ2, were found tobe associated with schizophrenia in 213 French Canadian cases and 24 1controls, with a modest repl icat ion in 183 Russian cases and 183 controls. Yeasttwo-hybrid experiments led to the identi f icat ion of o-amino acid oxidase (DAA?,12q2 ) interacting with G72, and a modest associat ion with schizophrenia wasdemonstrated for SNPs at DAAO. Physiological studies supported theplausibi l i tyof these genes as schizophrenia candidate genes.

'14 Shifman S, Bronstein M, Sternfeld M, et a/. A highly signif icant associat ion" between a COMT haplotype and schizophrenia. Am J Hum Genet 2002;

7 1:1296-13O2.New SNPs were identi f ied within COMT (22q11), a gene in the VCFS delet ionregion, and tested with pooled genotyping in approximately 700 Ashkenazi casesand 3000-5000 controls. Positive results were followed up with individualgenotyping, and a strong associat ion of schizophrenia was observed for severalSNP haplotypes.

Molecular genetics of schizophrenia Levinson 165

15 Mirnics K, Middleton FA, Stanwood GD, et a/. Disease-specif ic changes in' regulator of G-protein signal ing 4 (RGS4) expression in schizophrenia. Mol

Psychiatry 2001 ; 6:293-301.Using cDNA microarrays, the authors demonstrated that in brains of schizophreniasubjects, expression was consistently decreased for RGS4 (regulator of G-proteinsignal l ing ,1q21-q22), and not for other RGS genes or for 70 other genes in thiscytogenetic region. This is one of the first sophisticated microarray studies ofschizophrenia, and i t is l ikely that these studies wil l yield imporlant information thatwil l inform genetic studres.

16 Hemby SE, Ginsberg SD, Brunk B, ef a/. Gene expression profi le for" schizophrenia: discrete neuron transcription patterns in the entorhinal cortex.

Arch Gen Psychiatry 2002; 59:631-640.In one of the first applications of sophisticated microarray technologies toschizophrenia, these authors constructed an array of 18 000 mRNAs and studiedexpression in EC-l l stel late neurons from brains from eight schizophrenia cases(older patients with chronic i l lness, four on no medication) and nine controls. Themost prominent differences observed in a secondary more specific analysis werethe decreased expression of GRINl (NMDA receptor 1, 9q34,3), GNAIl [guaninenucleotide binding protein (G-protein) alpha inhibit ing activi ty polypeptide 1,7q211, SYP (synaptophysin, Xp1 1.23-p1 1.22), SNAP23 (synaptosomat-asso-ciated protein 23, 15q14) and SNAP21 Qapl2-p11.2), and the increasedexpression of GNG2 (G-protein gamma subunit 2, 14q21).

17 Chowdari KV, Mirnics K, Semwal P, et al. Associat ion and l inkage analyses" of RGS4 polymorphisms in schizophrenia. Hum Mol Genet 20Q2; 11:1373-

1 380.Mirnics etal. [15'] had demonstrated the reduced expression of RGS4 (1q23.1,regulator of G-protein signalling 4) in brains from schizophrenia subjects in anelegant series of microarray studies. In this report, SNPs were identified in theregion, and haplotype analysis revealed four SNPs with a modest association withschizophrenia in two independent samples with a trend in a third sample.

18 Waterwort DM, Bassett AS, Brzustowicz LM. Recent advances in the" genetics of schizophrenia. Cell Mol Life Sci 2002; 59:331-348.An excel lent review of schizophrenia diagnosis, epidemiology, and l inkage andassociat ion f indings.

19 Baron M. Genetics of schizophrenia and the new mil lennium: progress ano" pit fal ls. Am J Hum Genet 2001; 68:299-312.An excel lent review of schizophrenia l inkage and associat ion f indings, focusing onmethodological issues. The author argues for the use o{ larger pedigrees andmodel-based linkage analyses, but carefully reviews findings from all studydesigns.

20 Risch N. Linkage strategies for genetically complex diseases, with specialreference to psychiatr ic disorders. Gene Epidemiol 1990; 7:3-16.

21 Hauser ER, Boehnke M, Guo SW, Risch N. Affected sib-pair intervalmapping and exclusion for complex genetic traits: sampling considerations.Genet Ep idemio l 1996; 13 :1 1 7 -138.

22 Suarez BK, Hampe CL, van Eerdewegh PV. Problems of repl icat ing l inkageclaims in psychiatry. In: Gershon ES and Cloninger CR, editors. Geneticapproaches to mental disorders. Washington, DC: American PsychiatricPress ; 1994. pp .23-46.

23 Goring HH, Terwil l iger JD, Blangero J. Large upward bias in est imation oflocus-specif ic effects from genomewide scans. Am J Hum Genet 2001;69 :1 357-1 369.

24 Brzustowicz LM, Hodgkinson KA, Chow EW, et al. Location of a majorsusceptibi l i ty locus for famil ial schizophrenia on chromosome 1q21-q22.Science 2000; 288:678-682.

25 Blouin JL, Dombroski BA, Nath SK, et a/. Schizophrenia susceptibi l i ty loci onchromosomes 13q32,8p21. Nat Genet 1998; 2Q:7O-73.

26 Lander E, Kruglyak L. Genetic dissection of complex traits: guidel ines forinterpreting and report ing l inkage results. Nat Genet 1995; 11:241-247.

27 Straub RE, Maclean CJ, O'Neil l FA, et a/. A potential vulnerabi l i ty locus forschizophrenia on chromosome 6p24*22: evidence for genetic heterogeneity.Nat Genet 1995; 11:287-293.

28 Straub RE, MacLean CJ, O'Neil l FA, et a/. Support for a possibleschizophrenia vulnerabi l i ty locus in region 5q22-31 in lr ish famil ies. MolPsych ia t ry 1997; 2 :148-155.

29 Kendler KS, Maclean CJ, O'Neil l FA, et al. Evidence for a schizophreniavulnerabi l i ty locus on chromosome 8p in the lr ish Study of High-DensitySchizophrenia Famil ies. Am J Psychiatry 1996; 153:1534-1540.

30 Straub RE, Maclean CJ, Mart in RB, et al. A schizophrenia locus may beloca ted in reg ion 10p15-p1 1 . Am J Med Genet 1998; 81 :296-301.

31 Altmuller J, Palmer LJ, Fischer G, et al. Genomewide scans of complexhuman diseases: true l inkage is hard to f ind. Am J Hum Genet 2001; 69:936-950.

32 Ekelund J, Hovatta l , Parker A, ef a/. Chromosome 1 loci in Finnishsch izophren ia fami l ies . Hum Mol Genet 2001; 10 :1611-1617.

166 Schizophrenia

33 Coon H, Myles-Worsley M, Tiobech J, et al. Evidence for a chromosome2p13-14 schizophrenia susceptibi l i ty locus in famil ies from Palau, Micro-nes ia . Mo l Psych ia t ry 1998; 3 :521-527.

34 Maziade M, Roy MA, Rouil lard E, et al. A search for specif ic and common" susceptibi l i ty loci for schizophrenia and bipolar disorder: a l inkage study in 13

target chromosomes. Mol Psychiatry 2OO1; 6:684-693.Part ial genome scan data (focused on candidate regions) for 19 Ouebecoisepedigrees with 70 narrow and nine broad cases (as well as bipolar cases andsome purely bipolar pedigrees). Suggestive l inkage was observed on chromo-somes 6p22-24 and' l 8q.

35 Schizophrenia Linkage Collaborative Group for Chromosomes 3, 6 and 8.Addit ional support for schizophrenia l inkage on chromosomes 6, 8: amult icenter study. Am J Med Genet '1 996; 67:580-594.

36 Levinson DF, Holmans P, Straub RE, et a/. Mult icenter l inkage study ofschizophrenia candidate regions on chromosomes 5q, 6q, 10p and 13q:Schizophrenia Linkage Collaborative Group l l l . Am J Hum Genet 2O0O;67:652-663.

37 Levinson DF, Holmans PA, Laurent C, et al. No major schizophrenia locus' detected on chromosome 1q in a large mult icenter sample. Science 2002;

296:739-741.A large mult icenter sample (779 pedigrees from eight schizophrenia samples) didnot produce signif icant evidence for l inkage on chromosome 1q. The authorstyped a 5 cM map spanning proximal and distal l inkage f indings.

38 Sull ivan PF, Eaves Li, Kendler KS, Neale MC. Genetic case-controlassociat ion studies in neuropsychiatry. Arch Gen Psychiatry 2Q01; 58:1015-1024.

39 Schwab SG, Knapp M, Mondaboni S, et a/. Support for associat ion ofschizophrenia with genetic variat ion in the 6p22.3 gene, dysbindin, in sib-pairfamil ies with l inkage and in an addit ional sample of tr iad famil ies. Am J HumGenet 2003; 72 :185-1 90 .

40 Pulver AE, Nestadt G, Goldberg R, et a/. Psychotic i l lness in patientsdiagnosed with velo-cardio-facial syndrome and their relat ives. J Nerv MentD i s 1 9 9 4 ; 1 8 2 : 4 7 6 - 4 7 8 .

41 Karayiorgou M, Morris MA, Morrow B, et al. Schizophrenia susceptibi l i tyassociated with interstitial deletions of chromosome 22q11. Proc Natl AcadSc i USA 1995; 92 :761 2-7616.

42 Murphy KC, Jones RG, Grif f i ths E, etal. Chromosome 22q11 delet ions. Anunder-recognised cause of idiopathic learning disabi l i ty. Br J Psychiatry 1998;1 72 :1 8O-1 83 .

43 Brzustowicz LM, Honer WG, Chow EW, et al. Linkage of famil ialschizophrenia to chromosome 13q32. Am J Hum Genet 1999; 65:1096-1 1 0 3 .

44 Herken H, Erdal ME. Catechol-O-methyltransferase gene polymorphism in' schizophrenia: evidence for association between symptomatology and

prognos is . Psych ia t r Genet 2001; 11 :105-109.No association with schizophrenia was observed for the val/met polymorphism in129 cases and 165 controls from Turkey, although a modest associat ion with theseverity of illness was noted.

45 Norton N, Kirov G, Zammit S, et al. Schizophrenia and functional' polymorphisms in the MAOA and COMf genes: no evidence for association

or epistasis. Am J Med Genet 2002; 114:491-496.No association was observed between schizophrenia and the vali 58metpolymorphism in COMT in 346 Brit ish cases and 334 controls. The A287Gpolymorphism was also negative in half of the sample.

46 Liou YJ, Tsai SJ, Hong CJ, et a/. Associat ion analysis of a functional catechol-' o-methyltransferase gene polymorphism in schizophrenic patients in Taiwan.

Neuropsychobiology 2OO1 ; 43l.1 1-1 4.No associat ion was observed between schizoohrenia and the val158metpolymorphism in COMT in 198 Chinese cases and 188 controls, although amodest associat ion (P=0.005) was reported.

47 Semwal P, Prasad S, Bhatia T, et al. Family-based association studies ofmonoaminergic gene polymorphisms among North Indians with schizophre-nia. Mol Psychiatry 2OO1; 6:220-224.

48 de Chaldee M, Corbex M, Campion D, et al. No evidence for l inkage' between COMT and schizophrenia in a French populat ion. Psychiatry Res

2001 ; 102:87-90.One of the negative associat ion studies for older polymorphisms in COMf, in asmall mult iplex family sample, using family-based assocation and NPL analyses.

49 Weinberger DR, Egan MF, Berlol ino A, et a/. Prefrontal neurons and the' genetics of schizophrenia. Biol Psychiatry 2001; 50:825-844.A thoughtful review of animal and human studies of the physiology of the cognit iverole of prefrontal cortex, and of evidence that sequence variat ion in COMf is anexample of a polymorphism that direct ly inf luences this physiology through i tseffects on dopaminergic transmission.

50 Wong AH, Buckle CE, Van Tol HH. Polymorphisms in dopamine receptors:what do they tel l us? Eur J Pharmacol 2000; 410:183-203.

51 Dubertret C, Gorwood P, Gouya L, et al. Associat ion and excess of' transmission of a DRD2 haplotype in a sample of French schizophrenic

patients. Schizophr Res 2001 i 49:2Q3-212.A modest associat ion (P=0.02-0.05) is reporled by schizophrenia and twopolymorphisms in DRD2 (dopamine receptor 2, 11q22-q23) in 50 French casesversus controls and by transmission disequil ibr ium test (TDT) in the same cases,with a trend towards a later age of onset predicting the association. PreviousDRD2 studies are also reviewed.

52 Virgos C, Martorel l L, Valero J, et al. Associat ion study of schizophrenia with' polymorphisms at six candidate genes. Schizophr Res 2001; 49:65-7 1.ln 262 Spanish cases versus 278 controls, an associat ion at P=0.035 wasobserved aI DRD2, but not at five other loci (DRD3, 5HTR2A, Nf-3, BDNF orcNrfl.

53 Hori H, Ohmori O, Shinkai T, et al. Associat ion analysis between two' {unctional dopamine D2 receptor gene polymorphisms and schizophrenia.

Am J Med Genet 2001; 105:176-178.No associat ion was observed between schizophrenia and two polymorphisms inDRD2 in 241 Japanese cases versus 201 contrors.

54 Rybakowski JK, Borkowska A, Czerski PM, Hauser J. Dopamine D3 receptor' (DRD3) gene polymorphism is associated with the intensity of eye movement

disturbances in schizophrenic patients and healthy subjects. Mol Psychiatry2 O O 1 ; 6 : 7 1 8 - 7 2 4 .

An associat ion (P=0.001-0.05) was observed between the severity of eyetracking deficits in schizophrenia patients and the Ser9Gly polymorphism in DRD3.

55 Anney RJ, Rees Ml, Bryan E, et al. Characterisation, mutation detection, and' association analysis of alternative promoters and 5' UTRs of the human

dopamine D3 receptor gene in schizophrenia. Mol Psychiatry 2OO2; 7:493-502.

No association was observed between schizophrenia and polymorphisms inalternative promoter and S'UTR regions of DRD3, and the finding of excesshomozygosity of DRDS (dopamine 3 receptor,3q13.3) in schizophrenia(previously reported by this and other groups) was not observed.

56 Muir WJ, Thomson ML, McKeon P, et al. Markers close to the dopamine D5' receptor gene (DRDS) show significant association with schizophrenia but

no t b ipo la r d isorder . Am J Med Genet 2001; 105:152-158.Modest associations were observed between schizophrenia and a polymorphismin DRD5 (dopamine 5 receptor,4p16) (P=0.024) and in a nearby microsatel l i temarker, D45615 (P=0.001), although the two markers were not in l inkagedisequil ibr ium (150 kb region).

57 Wil l iams J, Spurlock G, McGuff in P, et a/. Associat ion between schizo-phrenia and T102C polymorphism of the 5-hydroxytryptamine type 2a-receptor gene. European Multicentre Association Study of Schizophrenia(EMASS) Group. Lancet 1996; 347:1 294-1296.

58 Wil l iams J, McGuff in P, Nothen M, Owen MJ. Meta-analysis of associat ionbetween the 5-HT2a receptor T1O2C polymorphism and schizophrenia.EMASS Collaborative Group. European Multicentre Association Study ofSchizophrenia. Lancet 1997;, 349:1221.

59 Chen RY, Sham P, Chen EY, et al. No associat ion between T102Cpolymorphism of serotonin-2A receptor gene and cl inical phenotypes ofChinese schizophrenic patients. Psychiatry Res 2001; 105:175-185.

60 Haider MZ, Zahid MA. No evidence for an association between the 5-hydroxytryptamine 5-HT2a receptor gene and schizophrenia in Kuwaiti Arabs.Psychiatry Clin Neurosci 2OO2; 56:465-467.

6'1 Liu W, Gu N, Feng G, et al. Tentative association of the serotonin transporterwith schizophrenia and unipolar depression but not with bipolar disorder inHan Chinese. Pharmacogenetics 1999; 9:491-495.

62 Tsai SJ, Ouyang WC, Hong CJ. Associat ion for serotonin transporter gene' variable number tandem repeat polymorphism and schizophrenic disorders.

Neuropsychobiology 2002; 45: 1 31 -1 33.Very modest replication (P=0.043) was reported for the association of theserotonin transporter gene (S-HII-) with schizophrenia in Han Chinese (1 14 casesversus 127 controls), the populat ion in which such an associat ion was previouslyreoorted.

63 Sanders AR, Cao O, Taylor J, et al. Genetic diversity of the human serotoninreceptor 1B (HfR1B) gene. Genomics 200.1 ;72 :1-14 .

64 Niesler B, Weiss B, Fischer C, et al. Serotonin receptor gene HTR3Avariants in schizophrenic and bipolar affective patients. Pharmacogenetics2 0 0 1 ; 1 1 : 2 1 - 2 7 .

65 Schiffer HH. Glutamate receptor genes: susceptibi l i ty factors in schizo-' phrenia and depressive disorders? Mol Neurobiol 2OO2; 25:191-212.A useful review of the evidence support ing a role for glutamatergic mechanisms inschizophrenia.

66 Begni S, Popoli M, Moraschi S, et a/. Associat ion between the ionotropic' glutamate receptor kainate 3 (GR/l(3) ser310ala polymorphism and

schizophrenia. Mol Psychiatry 2QO2; 7:416-418.A modest associat ion (P=0.0105) was reporled between schizophrenia and aser31Oala polymorphism in GR/KS (ionotropic glutamate receptor kainate 3,1p34-33) in 99 l tal ian cases and 116 controls.

67 Mart i SB, Cichon S, Propping P, Nothen M. Metabotropic glutamate receptor' 3 (GRM3) gene variat ion is not associated with schizophrenia or bipolar

affect ive disorder in the German populat ion. Am J Med Genet 2002; i14:46-50.

An association observed between schizophrenia and GRM3 (metabatropicglutamate receptor 3, 7q21 .1-q21.2) in 265 cases versus 227 controls(P=O.OO22) could not be confirmed in an addit ional 288 cases versus 162controls and 128 tr ios (German).

68 Ohtsuki T, Toru M, Arinami T. Mutation screening of the metabotropic' glutamate receptor mGluR4 (GRM4) gene in patients with schizophrenia.

Psych ia t r Genet 2001; 11 :79-83.No association was observed between schizophrenia and GRM4 (metabotropicglutamate receptor 4,6p21.2), using nine newly discovered sequence variants, inJapanese cases and controls.

69 Shibata H, Joo A, Fuj i i Y, et al. Associat ion study of polymorphisms in the' GluRb kainate receptor gene (GRIK1) with schizophrenia. Psychiatr Genet

2 0 0 1 ; 1 1 : 1 3 9 - 1 4 4 .No association was observed between schizophrenia and GR/KI (ionotropicglutamate receptor kainate 1,21q22.11) using three new and three previouslydescribed SNPs in approximately 200 Japanese cases and controls per analysis.

70 Joo A, Shibata H, Ninomiya H, et al. Structure and polymorphisms of the' human metabotropic glutamate receptor type 2 gene (GRM2): analysis of

associat ion with schizophrenia. Mol Psychiatry 2001; 6:186-192.No association was observed between schizophrenia and GRM2 (metabotropicglutamate receptor 2,3p12-p11) in 213 Japanese cases and 220 controls.

71 Blackwood DHR, Fordyce A, Walker MT, et a/. Schizophrenia and affectivedisorders - cosegregation with a translocation at chromosome 1q42 thatdirect ly disrupts brain-expressed genes: cl inical and p3O0 f indings in a family.Am J Hum Genet 2001; 69:428-433.

72 Mil lar iK, Christ ie S, Anderson S, et al. Genomic structure and local isat ion" within a linkage hotspot oI Disrupted In Schizophrenia 1, a gene disrupted by

a translocation segregating with schizophrenia. Mol Psychiatry 2001; 6:123-178.

This is the initial report on the genes DISCl and D/SC2 (parallel antisense genes)found in the 1q breakpo in t reg ion o f a (1 ;11) (q42.1 ; q l4 .3 ) t rans loca t ion tha tsegregates with schizophrenia and other psychiatric disorders in a large Scottishpeorgree.

73 Devon RS, Anderson S, Teague PW, et al. ldentification of polymorphisms' within Disrupted in Schizophrenia 1 and Disrupted in Schizophrenia 2, and an

investigation of their association with schizophrenia and bipolar affectivedisorder. Psychiatr Genet 2001; 11:71-78.

Two plausible candidate genes for schizophrenia were discovered in thebreakpoint of a (1;11) @a21i q14.3) translocation that segregates withschizophrenia and other psychiatric disorders in a previously described very largeScottish pedigree. Although association with schizophrenia was not observed forthe SNPs described here in these genes, these genes may be of importance giventhat this is one of the few relevant cytogenetic f indings in schizophrenia.

74 Semple CA, Devon RS, Le Hellard S, Porteous DJ. ldenti f icat ion of genes' from a schizophrenia-l inked translocation breakpoint region. Genomics 2OO1;

73:123-126.This paper describes the possible schizophrenia candidate genes in thebreakpoint region on chromosome 11, for the 1;11 breakpoint described in thisgroup's papers on the 1q breakpoint region. No associat ion data for schizophreniaare presented.

75 Maziade M, Fournier A, Phaneuf D, et al. Chromosome 1q12-g22 l inkage' results in eastern Ouebec families affected by schizophrenia. Am J Med

Genet 2OO2; 114:51-55.Linkage was not observed on proximal chromosome 1q in 21 large FrenchCanadian pedigrees.

76 lmamura A, Tsujita T, Kayashima T, et al. Lack of association between the' hKCa3 gene and Japanese schizophrenia patients. Psychiatr Genet 2001;

11:227-229.No association was observed between schizophrenia and a CAG reoeat inKCNN3 (1q21 .2) in 1 12 Japanese cases and 102 controls.

77 Ritsner M, Modai l , Ziv H, et a/. An associat ion of CAG repeats at the' KCNN3 locus with symptom dimensions of schizophrenia. Biol Psychiatry

2OO2; 51:788-794.Longer alleles of the CAG repeat in KCNN3 (1q21 .2) predicted a greater severityof negative symptoms and of paranoid symptoms and an earlier age at onset in117 lsrael i cases (P<0.001 overal l mult ivariate analvsis of variance).

Molecular genetics of schizophrenia Levinson 167

78 Boin F, Zanardini R, Piol i R, et al. Associat ion between G308A tumor' necrosis factor alpha gene polymorphism and schizophrenia. Mol Psychiatry

2 0 0 1 ; 6 : 7 9 - 8 2 .In 84 patients versus 138 controls, modest associat ion (P=0.0024) is reportedbetween schizophrenia and an SNP in TNF2A (tumor necrosis factor alpha) in6p21 .1-21.3, the most centromeric port ion of the rather broad candidate regionon 6p.

79 Wei J, Hemmings GP. The NOTCH4 locus is associated with susceptibi l i tyto schizophrenia. Nat Genet 2000; 25:376-377.

80 Sklar P, Schwab SG, Wil l iams NM, et al. Associat ion analysis of NOTCH4loci in schizophrenia using family and populat ion-based controls [JournalArt iclel. Nat Genet 2QO1; 28:126-128.

81 McGinnis RE, Fox H, Yates P, et al. Fai lure to confirm NOTCH4 associat ionwith schizophrenia in a large populat ion-based sample from Scotland. NatGenet 2001 ; 28:128-129.

82 Fan JB, Tang JX, Gu NF, et al. A family-based and case-control associat ion' study of the NOTCH4 gene and schizophrenia. Mol Psychiatry 2OO2;7:1OO-

1 0 3 .No association was observed between schizophrenia and NOfCH4 SNPs in 544Han Chinese cases and 621 controls, as well as an analysis of over 300 tr ios.

83 lmai K, Harada S, Kawanishi Y, et a/. The (CTG)n polymorphism in the' NOTCH4 gene is not associated with schizophrenia in Japanese individuals.

BMC Psych ia t ry 2001; 1 :1 .No association was observed between schizophrenia and NOfCH4 SNPs in 102Japanese cases versus 100 controls.

84 Ujike H, Takehisa Y, Takaki M, et al. NOTCH4 gene polymorphism andsusceptibility to schizophrenia and schizoaffective disorder [Journal Article].Neurosci Lett 2001 ; 3O1 :41-44.

85 Schwab SG, Hallmayer J, Freimann J, et al. Investigation of l inkage and' associat ion/l inkage disequil ibr ium of HLA A-, DQA|-, DAg-, and DRBl-

al leles in 69 sib-pair- and 89 tr io-famil ies with schizophrenia. Am J Med Genet2 O O 2 ; 1 1 4 : 3 1 5 - 3 2 0 .

In 69 sib-pair famil ies and 89 tr ios (mostly European, a few non-Ashkenazilsraelis), no significant association was observed between schizophrenia andDAil, DRBI, DAA| , or HLA-A al leles or haplotypes, although this groupcontinued to observe suggestive evidence of l inkage (P= 0.0004) peaking at HLA-DABICAR (6p21.2) .

86 Matsumoto S, Sasaki T, lmamura A, et al. HLA class I distr ibution in' Japanese patients with schizophrenia. Am J Med Genet 2OO2; 114:42-45.No association was observed between schizophrenia and any of 45 class I HLAalleles ( i .e. the lowest uncorrected P value was 0.01), in 98 Japanese casesversus 393 controls.

A7 Li T, Underhi l l J, Liu XH, et al. Transmission disequil ibr ium analysis of HLA' class l l DRBI, DAA| , DOB1 and DPB1 polymorphisms in schizophrenia

using family tr ios from a Han Chinese populat ion. Schizophr Res 2001;49:73-78.

In 165 Han Chinese proband-parent tr ios, no signif icant associat ion of HLADRB1, DAAI , DOB or DPB alleles or haplotypes was observed after correctionfor mult iple tesing (the lowest global Pvalue was 0.019).

88 Blaveri E, Kalsi G, Lawrence J, et al. Genetic association studies of' schizophrenia using the 8p21-22 genes: prepronociceptin (PNOC), neuronal

nicotinic cholinergic receptor alpha polypeptide 2 (CHRNA2) and arylamineN-acetyltransferase 1 (NAf1). Eur J Hum Genet 2001; 9:469-472.

No association with schizophrenia was observed for markers in three genes in theap21-22 region implicated by schizophrenia l inkage studies.

89 Freedman R, Coon H, Myles-Worsley M, et a/. Linkage of a neurophysio-logical deficit in schizophrenia to a chromosome 15 locus. Proc Natl Acad SciU S A 1 9 9 7 ; 9 4 : 5 8 7 - 5 9 2 .

90 Freedman R, Adams CE, Adler LE, et a/. Inhibitory neurophysiological defici tas a phenotype for genetic investigation of schizophrenia. Am J Med Genet2000; 97:58-64.

91 Freedman R, Adams CE, Leonard S. The alphaT-nicotinic acetylchol inereceptor and the pathology of hippocampal interneurons in schizophrenia. JChem Neuroanat 2000; 20:299-306.

92 Freedman R, Leonard S, Gault JM, et a/. Linkage disequil ibr ium forschizophrenia at the chromosome 15q13-14 locus of the alphaT-nicotinicacetylchol ine receptor subunit gene (CHRNA4. Am J Med Genet 200 1;1Q5:2Q-22.

93 Freedman R, Leonard S, Olincy A, et al. Evidence for the mult igenic" inheritance of schizophrenia. Am J Med Genet 2001; 105:794-800.The authors re-analysed al l famil ies from the NIMH Genetics Init iat ive schizo-phrenia project, combining European and Afr ican-American pedigrees (they hadpreviously been separately analysed) and using a parametric analysis under adominant transmission. Signif icant l inkage was observed on chromosome 15qnear the CHRNAT locus.

168 Schizophren ia

94 Liu CM, Hwu HG, Lin MW, et a/. Suggestive evidence for l inkage o{' schizophrenia to markers at chromosome 15q13-14 in Taiwanese famil ies.

Am J Med Genet 2001; 105:658-661.Suggestive evidence for linkage was observed to markers near CHRNAT(P=0.0003 and 0.0008 under broad and narrow diagnostic models by NPLanalysis) in 52 Taiwanese schizophrenia famil ies.

95 Tsuang DW, Skol AD, Faraone SV, et al. Veterans Affairs Cooperative Study.' Examination of genetic l inkage of chromosome . l 5 to schizophrenia in a large

Veterans Affairs Cooperative Study sample. Am J Med Genet 2O0 l;1 05:662-668.

Modest evidence for l inkage to markers near CHRNAZ on chromosome 1bq wasobserved in 166 schizophrenia famil ies (216 affected sib-pairs), with a maximumNPL score o f 1 .65 .

96 Gejman PV, Sanders AR, Badner )A, et al. Linkage analysis of schizophrenia' to chromosome 15. Am J Med Genet 2001; 105:789-793.Modest evidence for l inkage to markers near CHRNAZ on chromosome 15q wasobserved in 68 schizophrenia famil ies with a maximum lod score (GenehunterP lus) o f 2 .0 .

97 Xu J, Pato MT, Torre CD, et al. Evidence for l inkage disequil ibr ium between' the alpha 7-nicotinic receptor gene (CHRNAZ) locus and schizophrenia in

Azorean fami l ies . Am J Med Genet 2001; 105:669-674.An association between 9HRNAT and schizophrenia was observed in 31 familiesfrom the Azores, with P=o.ooo4 for one marker. Differences related to paternal lyand maternal ly transmitted al leles were also observed. The small sample size l imitsthe interpretat ion of these f indings.

98 Meyer J, Ortega G, Schraut K, et al. Exclusion of the neuronal nicotinic' acetylchol ine receptor alphaT subunit gene as a candidate for catatonic

schizophrenia in a large family support ing the chromosome 15q13-22 locus.Mol Psychiatry 2OO2; 7 :220-223.

These authors had observed l inkage between the region of 15q containingCHRNAT and an alternative phenotype (periodic catatonia), but assocation withmarkers in CHRNAT was not observed.

99 Arinami T, Ohtsuki T, Takase K, et al. Screening for 22q11 delet ions in a' schizophrenia populat ion. Schizophr Res 2001 ; 52:162-170.Of 300 Japanese schizophrenia patients, one had a 22q11.2 (VCFS) delet ion,with no obvious stigmata of VCFS.

100 De Luca A, Pasini A, Amati F, et al. Association study of a promoter' polymorphism of UFDlL gene with schizophrenia. Am J Med Genet 2001;

105:529-533.In BB patients and 92 controls from ltaly, evidence for the association ofschizophrenia wirh UFDlL was observed (P=0.009 {or the variant al lele), with amodest confirmation (P= 0.03) in 38 proband-parent tr ios from Canada. This is anubiquit in family gene in the VCFS delet ion region.

101 Saito T, Guan F, Papolos DF, et al. Polymorphism in SNAP29 gene promoter' region associated with schizophrenia. Mol Psychiatry 2001; 6:193-20' l .A modest association (P= 0.009) was observed for SNAP29 (a gene in the VCFSdelet ion region) and schizophrenia, but not bipolar disorder.

102 Takase K, Ohtsuki T, Migita O, et al. Associat ion of ZNF74 gene genotypes' w i th age-a t -onset o f sch izophren ia . Sch izophr Res 2001;52 :161- . |65 .Associat ion (P<0.0001) is reported here for age at onset of schizophrenia (butnot schizophrenia) with SNPs in the ZNF74 gene in 22q11.2 in the VCFS delet ionregion, in 300 Japanese cases versus 300 controls, and the finding was replicatedin a second sample of 169 schizophrenia patients (P= O.OOOl ), making this one ofthe few f indings repl icated in samples this large.

l03Feinstein C, El iez S, Blasey C, Reiss AL. Psychiatr ic disorders and' behavioral problems in chi ldren with velocardiofacial syndrome: usefulness

as phenotypic indicators of schizophrenia r isk. Biol Psychiatry 2002; 51 :312-3 1 B .

The authors compared 28 chi ldren with VCFS with 29 chi ldren with comparablecognit ive function, and found similar behavioral and psychiatr ic disorders. Theysuggest that the specif ici ty of the associat ion of schizophrenia with the VCFSdelet ion is unproved. However, they fai l to account {or the high rate ofschizophrenia in adults with VCFS, certainly higher than in other groups withmild mental retardation.

104 Meyer J, Huberth A, Ortega G, et a/. A missense mutation in a novel gene' encoding a pulative cation channel is associated with catatonic schizophrenia

in a large pedigree. Mol Psychiatry 200 1; 6:302-306.These authors had reported linkage of periodic catatonia (a phenotypic variant ofschizophrenia) to chromosome 22q13. Here they report the co-segregation of amissense mutation in WKL| (22q13.33) with the disorder in one extendedpedigree.

105 Devaney JM, Donarum EA, Brown KM, et a/. No missense mutation of WKL|' in a subgroup of probands with schizophrenia. Mol Psychiatry 2OO2; Z:419-

423.A negative study of the associat ion of schizophrenia with the WKL| gene on22q13.33, a region where l inkage had been reported to the periodic catatoniaphenotype.

l06Jorgensen TH, Borglum AD, Mors O, etal. Search for common haplotypes' on chromosome 22q in patients with schizophrenia or bipolar disorder from

the Faroe lslands. Am J Med Genet 2002; 114:245-252.This is a report on progress towards finding shared haplotypes on chromosome22q in schizophrenia cases from a populat ion isolate. No missense mutations werefound in WKL| , a gene in the region of interest in these famil ies.

107 McOuil l in A, Kalsi G, Moorey H, et al. A novel polymorphism in exon 1 1 of theWKL| gene, shows no associat ion with schizophrenia. Eur J Hum Genet2OO2; 10:491-494.

108Gross J, Grimm O, Ortega G, et a/. Mutational analysis of the neuronal' cadherin gene CELSRl and exclusion as a candidate for catatonic

schizophrenia in a large family. Psychiatr Genet 2001; 1 1:197-200.A negative association study in periodic catatonia (a phenotypic variant ofschizophrenia) tor a 22q13.3 gene.

109 Mlmmack ML, Ryan M, Baba H, et a/. Gene expression analysis in' schizophrenia: reproducible up-regulat ion of several members of the

apolipoprotein L family located in a high-susceptibi l i ty locus for schizophreniaon chromosome 22. Proc Natl Acad Sci U S A 2002; 99:4680-4685.

This is an intr iguing study. A custom-made microarray was used to screen brainsof schizophrenia and control subjects for dif ferences in expression. Upregulat ionwas detected for APOL| (apol ipoprotein L1) (2.6{old), a f inding that was thenrepl icated in a second brain col lect ion. APOL2 and APOL4 were then shown tobe similarly upregulated in two brain col lect ions. Al l o{ these loci are in 22q12, nearbut not in the VCFS region.

110 Tachikawa H, Harada S, Kawanishi Y, et a/. Linked polymorphisms' (- 333G > T and - 2BOA > G) in the promoter region of the CCK-A receptor

gene may be associated with schizophrenia. Psychiatry Res 2OO1; 103l.147-1 55 .

Associat ion is reporled between schizophrenia and l inked polymorphisms in theCCKAR promoter region.

111 Wang Z, Wassink T, Andreasen NC, Crowe RR. Possible associat ion of a' cholecystokinin promoter variant to schizophrenia. Am J Med Genet 2OO2;

1 1 4:479-482.Modest association (P<0.05 in several tests) was reported between schizo-phrenia and a polymorphism in the CCK promoter region, in 85 cases versus 247controls, and in 60 tr ios (from the same cases).

112 Hattori E, Yamada K, Toyota f , et al. Association studies of the CT repeatpolymorphism in the 5' upstream region of the cholecystokinin B receptorgene with panic disorder and schizophrenia in Japanese subjects. Am J MedGenet 2001 ; 105:779-782.

113 Yoshikawa T, Kikuchi M, Saito K, et al. Evidence for associat ion o{ the myo-' inositol monophosphatase 2 (\MPA2) gene with schizophrenia in Japanese

samples. Mol Psychiatry 2QO1 ; 6:202-210.Associat ion (P=0.031-0.000.|) was reported between schizophrenia and threeSNPs in myo-inositol monophosphatase 2 (\MPA2, 18p11 .2), in 302 Japaneseschizophrenia cases versus 308 controls and 205 affective disorder cases.

114Jun TY, Pae CU, Chae JH, eta/. Polymorphism of CTLA-4 gene at posit ion' 49 of exon I may be associated with schizophrenia in the Korean populat ion.

Psychiatry Res 2002; 1 1O:19-25.CTLA4 (cytotoxic T-lymphocyte antigen-4, 2q33) was studied in 1 16 cases versus149 controls from Korea, based on an immune hypothesis. Modest evidence forassociat ion was reported (P=0.003 for al leles).

l l5Ohtsuki T, Sakurai K, Dou H, et al. Mutation analysis of the NMDAR2B' (GRIN2B) gene in schizophrenia. Mol Psychiatry 2001 ;6:211-216.Although only modest evidence of association (P= 0.004 for homozygosity ofl inked SNPs) was reported between schizophrenia and GRIN2B (12p12), theNMDA receptor 28, the sample size of 268 Japanese cases and 337 controlsmakes the study of some interest in generating hypotheses.

116 Maes M, Delanghe J, Bocchio Chiavetto L, et a/. Haptoglobin polymorphism' and schizophrenia: genetic variat ion on chromosome 16. Psychiatry Res

2 0 0 1 ; 1 0 4 : 1 - 9 .Modest associat ion is reported between schizophrenia and Hp (haptoglobin,16q22.1), in 98 schizophrenia cases compared with established genotype andplasma level distr ibutions in this northern l tal ian populat ion. The authors present abrief but useful summary of evidence for involvement in the inflammatory system inschizophrenia; Hp is one of the acute phase inf lammatory reaction proteins.

117 Hong CJ, Tsai SJ, Wang YC. Associat ion between tryptophan hydroxylase' gene polymorphism (4218C) and schizophrenic disorders. Schizophr Res

2001; 49 :59-63.Associat ion between schizophrenia and the A218C polymorphism in TPH(tryptophan hydroxylase, 11p14-p15.3) was observed (P=0.002) in 196 HanChinese cases versus 251 controls. This is the rate-l imit ing enzyme for serotoninbiosynthesis.

118Chen CH, Hung CC, Wei FC, Koong FJ . Debr isoqu ine 4-hydroxyrase(CYP2DO) genetic polymorphisms and susceptibi l i ty to schizophrenia inCh inese pa t ien ts f rom Ta iwan. Psych ia t r Genet 2001; 11 :153-155.

119Chiu HJ, Wang YC, Chen JY, et al. Associat ion study of the p53-geneProT2Arg polymorphism in schizophrenia. Psychiatry Res 2OO1; 105:279-283.

120 Chowdari KV, Brandstaetter B, Semwal P, et al. Association studies olcytosol ic phospholipase A2 polymorphisms and schizophrenia among twoindependent family-based samples. Psychiatr Genet 2001 ; 11:2O7-212.

121 lshiguro H, Ohtsuki T, Okubo Y, et al. Associat ion analysis of the pituitaryadenyl cyclase activating peptide gene (PACAfl on chromosome 18p11 withschizophrenia and bipolar disorders. J Neural Transm - Gen Sect 2001;108:849-854.

122 lshiguro H, Okubo Y, Ohtsuki T, et al. Mutation analysis of the ret inoid Xreceptor beta, nuclear-related receptor 1, and peroxisome proliferator-activated receptor alpha genes in schizophrenia and alcohol dependence:possible haplotype association of nuclear-related receptor 1 gene to alcoholdependence. Am J Med Genet 2002; 114:15-23.

123Jun TY, Pae CU, Chae JH, et al. Report on lL-10 gene polymorphism atposit ion 819 for major depression and schizophrenia in Korean populat ion.Psychiatr Cl in Neurosci 2OA2; 56:1 77-180.

l24Chiavetto LB, Boin F, Zanardini R, et a/. Association between promoterpolymorphic haplotypes of interleukin-10 gene and schizophrenia. BiolPsychiatry 2AO2; 51 :480-484.

125 Kunugi H, Kato T, Fukuda R, et al. Association study of C825T polymorphismof the G-protein b3 subunit gene with schizophrenia and mood disorders. JNeural Transm - Gen Sect 2002; 109:213-218.

126 Lee K, Kunugi H, Nanko S. Glial cel l l ine-derived neurotrophic factor (GDNF)gene and schizophrenia: polymorphism screening and associat ion analysis.Psychiatr Res 2001; 1Q4:11-17.

127 Leroy S, Griffon N, Bourdel MC, et al. Schizophrenia and the cannabinoidreceptor type 1 (CB1): association study using a single-base polymorphismin coding exon 1. Am J Med Genet 2001 ; 105:749-752.

128 Matsumoto C, Ohmori O, Hori H, et al. Analysis of associalion between theGln192Arg polymorphism of the paraoxonase gene and schizophrenia inhumans. Neurosc i Le t t 2002; 321:165-168.

129Ouyang WC, Wang YC, Hong CJ, Tsai SJ. Estrogen receptor alpha genepolymorphism in schizophrenia: frequency, age at onset, symptomatology andprognosis. Psychiatr Genet 2001; 1 1:95-98.

l30Ventr igl ia M, Bocchio Chiavetto L, Bonvicini C, etal. Al lel ic variat ion in thehuman prodynorphin gene promoter and schizophrenia. Neuropsychobiology2OO2;46:17-21.

131 Segman RH, Shapira Y, Modai l , etal. Angiotensin convert ing enzyme geneinsertion/deletion polymorphism: case-control association studies in schizo-phrenia, major affective disorder, and tardive dyskinesia and a family-basedassociat ion study in schizophrenia. Am J Med Genet 2002; 114:31G-314.

'132 Zhang B, Tan Z, Zhang C, et al. Polymorphisms of chromogranin B geneassociated with schizophrenia in Chinese Han population. Neurosci Lett2OO2: 323:229-233.

133 Hattori M, Kunugi H, Akahane A, etal. Novel polymorphisms in the promoterregion of the neurotrophin-3 gene and their associations with schizophrenia.Am J Med Genet 2002; 114:304-309.

134 Leszczynska-Rodziewicz A, Czerski PM, Kapelski P, et a/. A polymorphism ofthe norepinephrine transporter gene in bipolar disorder and schizophrenia:lack of association. Neuropsychobiology 2OO2; 45:1 82-1 85.

135 Feng J, Zheng J, Gelernter J, et al. An in-frame delet ion in the alpha (2C)adrenergic receptor is common in Afr ican-Americans. Mol Psychiatry 2001;6 :1 68-1 72 .

136 Frieboes RM, Moises HW, Gattaz WF, et a/. Lack of association betweenschizophrenia and the phospholipase-A(2) genes IPLA2 and sPLA2. Am JMed Genet 2OQ1; 105:246-249.

137 Tuulio-Henriksson A, Haukka J, Partonen T, et al. Heritabi l i ty and number of' quantitative trait loci of neurocognitive functions in families with schizophrenia.

Am J Med Genet 2002; 114:483-490.Based on data from 264 schizophrenia patients and their relatives from agenetical ly isolated region in Finland, the authors suggest that measures ofworking memory show strong heritability, perhaps as a result of a very smallnumber of genetic loci, and could be used for mapping studies.

138 Cosway R, Byrne M, Clafferty R, et a/. Neuropsychological change in youngpeople at high risk for schizophrenia: results from the first two neuropsycho-logical assessments of the Edinburgh High Risk Study. Psychol Med 2OOO;3 0 : 1 1 1 1 - 1 1 2 1 .

139 Cosway R, Byrne M, Clafferly R, et a/. Sustained attention in young people at' high risk for schizophrenia. Psychol Med 2002; 32:277-286.Measures of attention were not found to differentiate significantly between therelatives of schizophrenia and control subjects.

Molecular genetics of schizophrenia Levinson 169

140 Steel RM, Whalley HC, Mil ler P, et al. Structural MRI of the brain in' presumed carriers of genes for schizophrenia, their affected and unaffected

sibl ings. J Neurol, Neurosurg Psychiatry 2OO2; 72:455458.Based on a unique set of analyses of 'obligate carriers' (well individuals who aresibl ings of a schizophrenia patient and who also have a schizophrenic o{fspring), areduced volume of the amygdalohippocampal complex is suggested to be aspecific correlate of genetic risk.

141 Narr KL, Cannon TD, Woods RP, et al. Genetic contr ibutions to altered' cal losal morphology in schizophrenia. J Neurosci 2OO2;22:3720-3729.An upward bowing of the corpus callosum was suggested to be a neuroanato-mical marker for vulnerability to schizophrenia in the co{wins of cases.

142Egan MF, Hyde TM, Bonomo JB, et al. Relat ive r isk of neurological signs in' sibl ings of patients with schizophrenia. Am J Psychiatry 2001; 158: 1827-

1 834.Neurological signs were found to be increased in the relatives of schizophreniapatients, but the low relative risk suggested that these signs might be poormarkers for genetic studies.

143 Scutt LE, Chow EW, Weksberg R, et a/. Patterns of dysmorphic features inschizophrenia. Am J Med Genet 2001 ; 105:713-723.

'144 Cardno AG, Sham PC, Farmer AE, et al. Heritability of Schneider's Jirst-rank' symptoms. Br J Psychiatry 2002; '1 80:35-38.First-rank psychotic symptoms were shown to be heritable in twins, although lessso than the categorical diagnosis of schizophrenia.

145 Cardno AG, Rijsdi jk FV, Sham PC, et al. A twin study of genetic relat ionships' between psychotic symptoms [see Comments]. Am J Psychiatry 2002;

159:539-545.Using the Maudsley twin series, the authors suggest that if schizophrenic, manicand depressive syndromes are assigned non-hierarchically, they show substantialoverlap in their genetic et iology. See comment by Kendler [146.].

146 Kendler KS. Hierarchy and heritabi l i ty: the role of diagnosis and modeling in' psychiatr ic genetics. Am J Psychiatry 2OO2;159:515-518.Commenting on Cardno et al. [144'1, the author points out that their data are moreconsistent with genetic factors influencing mood syndromes in individuals withschizophrenia, rather than for genetic overlap in the etiology of schizophrenia andbipolar disorder.

147 Brown AS, Schaefer CA, Wyatt RJ, et al. Paternal age and risk of" schizophrenia in adult offspring. Am J Psychiatry 2OQ2; 159:1528-1533.Using a birth cohort from which individuals with schizophrenia spectrum disordershad previously been ascertained, the authors demonstrate a small but significantrelationship between advancing paternal age and the risk of these disorders, aftercontrolling for confounding variables such as maternal age. This suggests thepossibility of de-novo genetic mutations in older fathers as factors in schizophrenia.

148 Mafaspina D, Corcoran C, Fahim C, et al. Paternal age and sporadic' schizophrenia: evidence for de novo mutations. Am J Med Genet 2002;

1 1 4:299-303.Schizophrenia patients without a family history had significantly older fathers thanthose with a family history, supporting the hypothesis of de-novo genetic mutationsin older fathers.

149 Malaspina D. Paternal factors and schizophrenia risk: de novo mutations and' imprint ing. Schizophr Bull 2001; 27:379-393.The author reviews neurodevelopmental disorders previously associated with de-novo germline mutations in older fathers, and discusses mechanisms by which thiseffect can occur, relating it to data for schizophrenia.

150 Lewis DA, Levitt P. Schizophrenia as a disorder of neurodevelopment" [Review]. [1 18 re{s] Annu Rev Neuroscience 2QO2;25:409-432.This is a comprehensive review of evidence favoring a neurodevelopmentaletiology ('pathogenetic biological events are present much earlier in life than theonset of the features of the i l lness', p.411). Evidence for such inf luences arereviewed by time period (prenatal, perinatal, childhood and adolescence).Emphasis is placed on the strength of evidence for a relationship betweenschizophrenia and complications during labor and del ivery, and the presence ofsubtle behaviors that predate active illness by many years. A polygenic-multifactorial model is favored whereby genetic predisposition which altersdevelopment, gene-environment interactions that influence or trigger the geneticeffects, and cumulative effects over time of the altered development, leading to acomparably siable state that is relatively difficult to alter.

151 Maynard TM, Sikich L, Lieberman JA, LaMantia AS. Neural development,' cel l-cel l signal ing, and the two-hit hypothesis of schizophrenia. Schizophr

Bu l l 2001: 273:457-476These authors argue for a 'two-hit' hypothesis of schizophrenia whereby genetic orenvironmental factors alter early development, and then a second factor closer tothe time of onset triggers the actual disease.

17O Schizophrenia

152 Schiffman J, Ekstrom M, LaBrie J, et al. Minor physical anomalies and" schizophrenia spectrum disorders: a prospective investigation. Am J

Psychiatry 2002; 1 59:238-243.In this unique study, 265 chi ldren from a 1959-1961 Danish birth cohort (90 with aparent with schizophrenia, 93 with a parent with some other psychiatric disorder,and 82 with no record of a psychiatrically ill parent) had been rated for minorphysical anomalies at 11-13 years of age. Psychiatr ic diagnosis was thendetermined at 31-33 years by direct interview or records. of 131 chi ldren withthree or more anomalies, 20 (15.3%) developed a schizophrenia spectrumdisorder, compared with six out of 1 1 1 chi ldren with two or fewer anomalies(5.4%) (P<0.01). For the high-r isk group the proport ions were 12 out of 39(30. 1%) versus f ive out of 42 (11.9%) (P<0.04). These are str iking resultssuggesting early, physical-developmental effects of genetic factors, although thehigh rate of schizophrenia spectrum disorders in this study populat ion is a bitsurprising.

153 Bassett AS, Chow EW, O'Neil l S, Brzustowicz LM. Genetic insights into the' neurodevelopmental hypothesis of schizophrenia. Schizophr Bull 2001;

27 :417-43O.This article provides a useful review of neurodevelopmental research inschizophrenia from a group that has focused on the study of classic geneticfeatures such as physical anomalies and the VCFS syndrome.

154Delisi LE. speech disorder in schizophrenia: review of the l i terature and' exploration of its relation to the uniquely human capacity for language.

Schizophr Bull 2001 ; 27:481-496.Based on data from a study oJ structured ratings of elements of language in free-form speech, the author proposes that the pathology of schizophrenia may berelated to genes underlying uniquely human characterist ics of language.

' f 55 Judson R, Sal isbury B, Schneider J, et al. How many SNPs does a genome-wide haplotype map require? Pharmacogenomics 2002; 3:379-391.

156 Gabriel SB, Schaffner SF, Nguyen H, et al. The structure of haplotype blocksin the human genome. Science 2OO2;296:2225-2229.

157 Daly MJ, Rioux JD, Schaffner SF, et al. High-resolut ion haplotype structure inthe human genome. Nat Genet 2OO1;29:229-232.

158 Bouchie A. Haplotype map planned. Nat Biotechnol 2001; 19:704.

159 Weiss KM, Terwil l iger JD. How many diseases does i t take to map a genewi th SNPs? Nat Genet 2000; 26 :151-157.

' , | 60 Faham M, Baharloo S, Tomitaka S, ef al. Mismatch repair detection (MRD):high{hroughput scanning for DNA variat ions. Hum Mol Genet 200.1 ;1 0 :1 657-1 664.

161 Conrad P. Genetic optimism: framing genes and mental i l lness in the news.' Cult Med Psychiatry 2001;25:225-247.The author analyzes news reports about psychiatric genetics findings and critiquesthe 'genetic optimism' that emphasizes the inevitabi l i ty of discovering genes andthe good outcomes of those discoveries, whereas negative or retracted findingsreceive no attention. Scientific colleagues in genetics and other fields are prone tosimilar distort ions, as discussed in the present art icle.