Genome-wide linkage analysis of ADHD using high- density SNP ...

Molecular Psychiatry (2008) 13, 522–530 

& 2008 Nature Publishing Group All rights reserved 1359-4184/08 $30.00 

www.nature.com/mp 

ORIGINAL ARTICLE 

Genome-wide linkage analysis of ADHD using highdensity 

SNP arrays: novel loci at 5q13.1 and 14q12 

M Romanos 1,7 , C Freitag 2,7 , C Jacob 3,7 , DW Craig 4 , A Dempfle 5 , TT Nguyen 5 , R Halperin 4 , S Walitza 1 , 

TJ Renner 1 , C Seitz 2 , J Romanos 3 , H Palmason 6 , A Reif 3 , M Heine 3 , C Windemuth-Kieselbach 5 , 

C Vogler 6 , J Sigmund 6 , A Warnke 1 , H Schäfer 5 , J Meyer 6 , DA Stephan 4 and KP Lesch 3 

1 ADHD Clinical Research Program, Department of Child and Adolescent Psychiatry and Psychotherapy, University of 

Wuerzburg, Wuerzburg, Germany; 2 Department of Child and Adolescent Psychiatry, Saarland University Hospital, Homburg, 

Germany; 3 ADHD Clinical Research Program, Department of Psychiatry and Psychotherapy, University of Wuerzburg, 

Wuerzburg, Germany; 4 Neurogenomics Division, Translational Genomics Research Institute, Phoenix, AZ, USA; 5 Institute of 

Medical Biometry and Epidemiology, University of Marburg, Marburg, Germany and 6 Department of Neurobehavioral Genetics, 

University of Trier, Trier, Germany 

Introduction 

Previous genome-wide linkage studies applied the affected sib-pair design; one investigated 

extended pedigrees of a genetic isolate. Here, results of a genome-wide high-density linkage 

scan of attention-deficit/hyperactivity disorder (ADHD) using an array-based genotyping of 

B50 K single nucleotide polymorphism (SNPs) markers are presented. We investigated eight 

extended pedigrees of German origin that were non-related, not part of a genetic isolate and 

ascertained on the basis of clinical referral. Two parametric analyses maximizing LOD scores 

(MOD) and a non-parametric analysis for both a broad and a narrow phenotype approach were 

conducted. Novel linkage loci across all families were detected at 2q35, 5q13.1, 6q22-23 and 

14q12, within individual families at 18q11.2-12.3. Further linkage regions at 7q21.11, 9q22 

and 16q24.1 in all families, and at 1q25.1, 1q25.3, 9q31.1-33.1, 9q33, 12p13.33, 15q11.2-13.3 and 

16p12.3-12.2 in individual families replicate previous findings. High-resolution linkage 

mapping points to several novel candidate genes characterized by dense expression in the 

brain and potential impact on disorder-relevant synaptic transmission. Our study provides 

further evidence for common gene effects throughout different populations despite the 

complex multifactorial etiology of ADHD. 

Molecular Psychiatry (2008) 13, 522–530; doi:10.1038/mp.2008.12; published online 26 February 2008 

Keywords: attention-deficit/hyperactivity disorder; ADHD; genome-wide scan; linkage; 

pedigrees; 50 K SNP array 

Attention-deficit/hyperactivity disorder (ADHD; MIM 

no. 143465) is a categorical classification of a complex 

behavioral phenotype comprising inattention, motor 

hyperactivity and impulsivity. The etiologically heterogeneous 

syndrome represents the most common 

child and adolescent behavioral disorder with similar 

prevalence rates throughout different cultural settings 

and independent of the applied classification 

system. 1 ADHD is increasingly being recognized as 

highly persistent into adulthood associated with 

Correspondence: Dr M Romanos, ADHD Clinical Research 

Program, Department of Child and Adolescent Psychiatry, University 

of Wuerzburg, Füchsleinstr., Wuerzburg 97080, Germany. 

E-mail: Romanos@kjp.uni-wuerzburg.de 

or Professor Dr KP Lesch, ADHD Clinical Research Program, 

Department of Psychiatry and Psychotherapy, University of 

Wuerzburg, Füchsleinstr., Wuerzburg 97080, Germany. 

E-mail: kplesch@mail.uni-wuerzburg.de 

7These authors contributed equally to the study. 

Received 26 August 2007; revised 17 January 2008; accepted 17 

January 2008; published online 26 February 2008 

considerable risk for psychiatric comorbidity and 

failure in psychosocial adaptation. 2–4 Although estimates 

from twin studies have consistently shown a 

heritability of 70–80%, 5 the underlying molecular 

genetic mechanisms remain to be elucidated. Association 

studies reported inconsistent results on 

variants in dopaminergic, serotoninergic and synaptosomal 

genes. 6–11 Genome-wide linkage scans revealed 

several susceptibility loci: four studies employed an 

affected sib-pair design, 7,8,12,13 one investigated 

extended pedigrees of a genetic isolate. 14 Linkage 

regions overlapping between the respective studies 

were identified across the genome; however, most 

findings did not reach the statistical significance 

threshold. Here, we report results from a linkage 

analysis on eight extended families densely segregating 

for ADHD using a B50 K single nucleotide 

polymorphism (SNP) array-based genotyping assay. 

Linkage analyses in extended families for a complex 

trait are generally based on two theoretical assumptions: 

(1) multiple small gene effects will contribute to 

the phenotype within the general population and thus

should be detectable in all or several families. 

(2) Within each pedigree, diminished genetic heterogeneity 

can be assumed 15 ; thus, strong family-specific 

gene effects are also likely to be found, which may or 

may not be extrapolated to general populations with 

ADHD. Family members were classified into a 

category of four, that is, definite ADHD; subclinical 

ADHD symptoms; no ADHD; unknown. For analysis 

of the narrow phenotype, subclinically affected 

individuals were classified as not affected, whereas 

for analysis of the broad phenotype they were 

classified as affected. We applied both parametric 

analysis methods maximizing logarithm of the odds 

ratio (LOD) scores for chromosomal sections (B500 

SNPs; MODglobal) or for single marker position 

(MODsingle) and a non-parametric approach (NPL). 

Methods 

Ascertainment and assessment 

We ascertained families of German origin through 

index children referred to three outpatient clinics in 

Wuerzburg, Homburg or Trier. The structure of the 

eight families (Pedigree 1–8; P1–P8) is shown in 

Figure 1 and summarized in Table 1. For the index 

child, strict inclusion and exclusion criteria were 

applied. Included index children were aged 6 or 

above, meeting criteria for the ADHD combined type 

according to DSM-IV. Index children that had a birth 

weight > 2000 g did not show any neurological or 

severe somatic disorder, drug abuse or autistic 

Genome-wide linkage analysis of ADHD 

M Romanos et al 

disorder and did not receive medication with central 

nervous effect (with the exception of methylphenidate). 

IQ was > 80. P1, P2, P3 and P4 were recruited 

in Wuerzburg; P5, P6, P7 and P8 in Homburg and 

Trier. When parents reported individuals in the 

extended family with presumable or definite affection 

with ADHD, pedigrees were drawn to determine 

family size and structure. Reported ADHD symptoms 

in more than two generations resulted in intensified 

recruitment of further family members either by 

invitation to our ADHD family outpatient clinic or 

by home visits. All participants signed informed 

written consent. The study was approved by the 

Ethics Committees of the Julius-Maximilians-University 

of Wuerzburg and the Saarland Doctors’ Association 

(SaarlÄndische Ärztekammer), respectively. 

Bilineality was not an exclusion criterion for 

recruitment, since it was presumably present in most 

recruited families. While bilineal families were 

excluded from ascertainment in the study by Arcos- 

Burgos et al., 14 a similar approach was not considered 

useful, since intra-familiar heterogeneity cannot be 

completely ruled out in complex traits. We recruited 

extended families that were non-related bearing in 

mind the possibility that different major gene effects 

might exist within the respective families. 

Children and adolescents were clinically characterized 

by a semi-structured interview (Kiddie-Sads-PL 

German version 16 or Kinder-DIPS 17 ), child behavior 

checklist (CBCL) 18 and by an ADHD diagnostic 

checklist applying DSM-IV criteria categorically and 

Figure 1 Pedigree structure of families. Individuals with a diagnosis of attention-deficit/hyperactivity disorder (ADHD) are 

depicted by black, the subclinical phenotype is represented by gray symbols. Unaffected family members are shown by white 

symbols, and individuals with unknown ADHD status are marked with a black circle in the symbol. A black dot beneath an 

individual indicates that DNA was available for genetic evaluation. Pedigrees are modified to preserve confidentiality. 

523 

Molecular Psychiatry

524 

Table 1 Sample description: number of individuals in multigenerational families according to the four diagnostic categories 

and number of DNA missing 

Family ADHD Subclinical No ADHD Unknown Total DNA missing 

P1 11 11 10 3 35 5 

P2 17 6 13 8 44 8 

P3 16 0 4 5 25 3 

P4 8 0 2 1 11 0 

P5 6 4 11 7 28 12 

P6 4 0 3 2 9 2 

P7 4 0 2 5 11 4 

P8 6 2 20 0 28 2 

dimensionally (ADHD-DC 19 ). Teacher ratings were 

obtained (TRF 18 or FBB-HKS 20 ). Psychosocial impairment 

was measured by self-rating for work/school 

situation, social and family life (Sheehan disability 

scale; SDS); 21 affective and psychosocial status was 

further determined by a German depression inventory 

(Depressionsinventar für Kinder und Jugendliche). 22 

Intelligence was measured by non-verbal CFT1/20 23,24 

in children or by MWT 25 in adults. Adult participants 

were also rated by the SDS and the ADHD-DC, 

based on a semistructured interview, thus enabling 

transgenerational comparability of current ADHD 

affection. Current and retrospective self-ratings 

were obtained. Symptoms were assessed by the 

Wender Utah Rating Scale (WURS-K), 26 retrospectively. 

In families recruited in Homburg and Trier, 

additionally the Wender Reimherr Interview 27 was 

used. A large number of the adult family members 

affected with ADHD were thoroughly examined by 

diagnostic interview for comorbidity on axis I and II 

(SKID-I, SKID-II). 28,29 

All family members were assessed by at least two 

clinicians experienced in diagnosis of childhood and 

adult ADHD (CF, CJ, HP, TR, JR, MR, CS, JS, CV, SW). 

In cases when assessment was incomplete, further 

information on the individual’s history was obtained 

through close relatives (in general spouse or parents). 

After recruitment, all participating individuals were 

assigned to one of four groups: (1) certain diagnosis of 

ADHD: fulfilling DSM-IV criteria of ADHD (six or 

more in at least one scale) according to ADHD-DC 

including rating of pervasiveness, WURS-K above 

cutoff (score above 30) in adults, psychosocial 

impairment present according to SDS (moderate, 

markedly or extreme impairment in more than one 

measure), typical childhood history; (2) subclinical 

affection with ADHD: four or five DSM-IV criteria for 

ADHD symptoms in a scale according to the ADHD- 

DC, WURS-K below threshold, no or mild psychosocial 

impairment according to SDS, inconsistent 

information about childhood history; (3) no affection 

with ADHD: 0–3 DSM-IV criteria according to ADHD- 

DC, no typical childhood history, no psychosocial 

Molecular Psychiatry 



Total 72 23 65 31 191 36 

Abbreviation: ADHD, attention-deficit/hyperactivity disorder. 

impairment; (4) unknown. Assignation to the respective 

category (1–4) was discussed and met by 

consensus of an expert committee of four in 

Wuerzburg (CJ, MR, JR, SW) or committee of three 

in Homburg (CF, HP, CS) considering all obtained 

information about the respective participant. In single 

cases, the committee met diagnosis ‘definite ADHD’ 

despite a WURS-K score below cutoff, when 

anamnestic information, biographic data and school 

reports clearly contradicted the self-ratings. Likewise, 

inaccurate parents’ reports on their children were 

adjusted when clinical impression and teacher reports 

gave clear indication for ADHD-symptoms of the 

child. Therefore, the rating in the ADHD-DC 

combined information from most applied instruments. 

Information on comorbid disorders diagnosed 

by interviews was relevant to discriminate phenocopies 

from ADHD-symptoms. Family members were 

classified as ‘unknown’, when definite assignation 

was not possible, reasons for this being lack of 

information due to non-compliance or unclear phenotype. 

Individuals having ADHD-symptoms without 

typical childhood history were considered as 

‘unknown’ as well as individuals with unclear 

etiology of symptoms (for example, chronic substance 

abuse, low birth weight < 2000 g, autistic symptoms). 

Genetic analysis 

Genomic DNA was extracted from blood samples 

according to standard protocols. DNA samples were 

diluted to a concentration of 50 ng ml 1 suspended in 

TE-buffer. For genotyping of B50 000 SNPs, the 

GeneChip Human Mapping 50 K Array Hind240 

(dBSNP build 125) was used according to the 

recommendations of Affymetrix. Initially 250 ng of 

genomic DNA was digested by HindIII. Ligation was 

performed by adaptors recognizing overhanging 

nucleotides (HindIII adaptor, Affymetrix Inc., Santa 

Clara, CA, USA) and T4 DNA ligase. Subsequently, 

adaptor-marked fragments were amplified by a 

generic primer. Purity of PCR was ensured using the 

QUIAquick PCR purification kit (QIAGEN, Hilden, 

Germany) according to the manufacturer’s protocol.

Purified amplicons were fragmented by DNase I and 

labeled by GeneChip Labeling Reagent (Affymetrix). 

Subsequently, 50 K arrays were used for hybridization. 

After washing in a Fluidics Station 450, the 

microarrays were stained in a three-step process with 

streptavidin phycoerythrin, biotinylated anti-streptavidin 

and again streptavidin phycoerythrin. Scanning 

was performed by a GeneChip Scanner 3000. Genotyping 

data were validated using GeneChipGCOS 

software. 

Statistical analysis 

The total genotype data of 57 244 SNPs (1151 

X-linked) and the corresponding map information 

from the GeneChip Human Mapping 50 K Array 

Hind240 (dBSNP build 125) were read into the 

program ALOHOMORA (v0.29, http://gmc.mdcberlin.de/alohomora), 

30 which was used to estimate 

allele frequencies and to check data for gender errors, 

Mendelian errors, deviation from Hardy–Weinberg 

equilibrium (HWE) in founders. Additionally, the 

data were checked for unlikely individual genotypes, 

which contradict information about gene flow 

provided by all other relatives (package Merlin 

v1.0-alpha, http://www.sph.umich.edu/csg/abecasis/ 

Merlin/index.html). 31 After the first control step, we 

excluded 30 478 SNPs with unknown position (385 

SNPs), minor allele frequency < 0.1 (24 541 SNPs), 

call rate < 0.8 (5679 SNPs), HWE w 2 > 12 (5275 SNPs) 

and with Mendelian errors (202 SNPs) or unlikely 

genotypes occurring in more than four families (eight 

SNPs). To avoid linkage peaks, which could arise 

from the artifact of linkage disequilibrium between 

the markers, we clustered the SNPs into haplotype 

blocks (defined by pairwise r 2 > 0.1) and selected from 

each block only one marker with the highest heterozygosity. 

A total of 10 061 autosomal SNPs (median 

(25–75th percentile), that is, distance 0.15 (0.05– 

0.37) cM, multipoint information content over all 

families 0.92 (0.91–0.93)) passed this second control 

step and were included in linkage analyses. The 

information content suggested by Kruglyak and 

colleagues 32 is a measurement for certainty of inheritance 

patterns provided by data at a given point 

on the genome. The described selection procedure not 

only guarantees reliable data quality, but it also still 

preserves, due to the dense marker map, nearly 

perfect information content. Additionally, the 

reduced marker set can be more easily handled in 

terms of computational demands. 

The ADHD status was defined narrowly or broadly 

according to the criteria mentioned above. We 

performed both parametric and non-parametric linkage 

(NPL) analyses in each family and in all families 

together using the program GENEHUNTER-MOD- 

SCORE (http://www.staff.uni-marburg.de/~strauchk/ 

software.html). The implemented procedure maximizes 

parametric LOD (MOD) scores with respect to 

the disease-allele frequency and three penetrances. 33 

We used the ‘modcalc global’ option (MODglobal) to 

assess MOD scores under the assumption that linkage 



signals within a certain chromosomal section might 

follow a uniform inheritance pattern, because they 

may arise from one and the same underlying gene 

variant. On the other hand, for a complex trait such as 

ADHD, it is possible that several causal gene variants 

with different inheritance patterns exist in close 

proximity to each other. To capture this case, we 

used the ‘modcalc single’ option (MODsingle) 

maximizing for each genetic position assumed for 

the putative disease locus. 

Linkage peaks with MOD score > 3.3 (according to 

the standard threshold proposed by Lander and 

Kruglyak 34 for global significant LOD score results) 

or NPL score > 5 (according to pointwise 

Pp2.9 10 7 ) were reported. A MOD score > 3 is 

required for linkage peaks, which overlap with 

findings of other studies. A minimal critical linkage 

region (MCR) around a MOD or NPL peak was defined 

(closest surrounding SNPs that showed < 1 MOD or 

NPL value below the maximum peak) to narrow a 

region that displays the highest probability to harbor 

true linkage. 

Assessment of the distribution for MOD score, due 

to its specific nature, requires complex simulation 

process, which has substantial computational demands 

in the situation of larger pedigrees. In addition, 

with respect to strong dependence between the 

considered combinations of phenotypes, families and 

methods, Bonferroni adjustment for multiple testing 

may imply conservativeness and loss of power. 

Therefore, we provide here only nominal results. 

Results 

Results are reported with reference to the guidelines 

for genome-wide linkage analyses of genetically 

complex traits suggested by Lander and Kruglyak. 34 

Linkage across all families was detected at 2q35, 

5q13.1, 6q22-23, 7q21.11, 9q22, 14q12, 16q24.1 (Table 

2). MOD scores of parametric analyses under the 

narrow and broad phenotype assumption are shown 

in Figures 2 and 3. Linkage in individual families was 

identified at 1q25.1, 1q25.3, 9q31.1-33.1, 9q33, 

12p13.33, 15q11.2-13.3, 16p12.3-12.2 and 18q11.2- 

12.3 (Table 3). 

Discussion 

Novel chromosomal loci across all families were 

predominately detected by parametric analysis of 

the narrow phenotype at 2q35, 5q13.1, 6q22.32- 

6q23.2 and 14q12 (Figures 2 and 3; Table 2). Overlap 

with loci reported previously was found at 7q21.11, 

9q22.1-9q22.2 and 16q24.1. In particular, we have 

identified a novel locus on chromosome 5q13.1 

showing linkage across all families with strong 

support from two individual families (P1 and P3). 

Contrasting with our findings, linkage on chromosome 

5 was reported at 5p13. Fine mapping suggested 

an association with a predisposing haplotype of 

the dopamine transporter gene (SLC6A3) located at 

525 


526 

Table 2 Linkage results in all families 


Locus Info Narrow Broad MCR DI fam n/r 

MODglobal MODsingle NPL MODglobal MODsingle NPL Marker Physical position (in Mb) 

2q35 0.92 2.98 3.40 2.32 1.20 1.26 2.26 rs1110998-rs1364637 217.17–218.53 0.94 P1, 2, 3, 8 n 

5q13.1 0.91 2.82 2.82 2.96 4.16 4.08 2.49 rs10515057 67.21 0.97 P1, 3, 8 n 

6q22-23 0.91 2.75 3.32 1.78 1.44 1.63 1.47 rs853966-rs9321354 127.08–132.95 0.58 P1, 3, 5, 6, 7, 8 n 

7q21.11 0.91 3.14 3.14 1.37 2.18 2.19 1.26 rs321983-rs929381 77.95–81.40 0.43 P1, 2, 3, 5, 6, 7, 8 r 

9q22 0.91 1.39 1.94 3.10 3.30 3.07 2.09 rs9314663-rs1777035 88.77–90.20 0.95 P1, 3, 4, 5, 6, 7, 8 r 

14q12 0.91 4.17 4.50 2.07 2.99 2.77 2.00 rs10483286-rs9323659 24.27–24.98 1.00 P1, 2, 3, 4, 7, 8 n 

16q24.1 0.90 3.26 3.20 3.57 1.88 2.03 3.86 rs7193075-rs3751797 82.85–85.12 0.78 P2, 3, 5, 7, 8 r 



Abbreviations: DI, dominance index of highest parametric score: near 0, additive model, near –1, recessive model, near þ 1, dominant model; fam, families contributing 

to score (family score > 0); Info, average information content of the respective chromosomal area; MCR, minimal critical region according to model of underlined score 

(SNP localization, 50 K Affymetrix chip Hind 240, dBSNP build 125); NPL, non-parametric linkage; n/r, novel/replication. 

Highest scores fulfilling criteria of Lander and Kruglyak (novel loci: MOD > 3.3 or NPL > 5.0; replication: MOD > 3.0) are underlined. 

1.45–1.5 Mb. 12 In our sample, fine mapping of 5p13 

revealed no segregation of marker haplotypes with 

ADHD thus implicating that these loci do not 

contribute to a large extent to the disorder in the 

investigated families. However, effects might still 

exist in the regions that are too small to be detected 

in our study due to lack of power. 

The locus identified on chromosome 6q22.32-23.2 

is close to but not overlapping with a region showing 

linkage to comorbid oppositional defiant disorder 

(ODD). 7,35 Because of the stochastic variance in 

linkage analysis, however, we cannot completely rule 

out that both linkage loci can be referred to the same 

interval. 36 No overlap was found between previously 

reported regions and the novel loci at 2q35 and 14q12. 

In contrast, linkage at 7q21.11 is a replication of a 

previous finding at 7q. 12 Apart from the linkage signal 

across all families on chromosome 9q22.1-22.2, two 

additional loci on chromosome 9q in two distinct 

families (P1: 9q31.1-33.1; P5: 9q33.2-33.3) were 

identified, the latter showing the highest NPL score 

(9.51; broad phenotype) detected in this study. These 

regions overlap partially with known ADHD-related 

loci 12,37,38 and with a locus that showed linkage to 

comorbid conduct disorder. 35 It is noteworthy that the 

linkage peak of the recent linkage publication by 

Asherson and coworkers 39 is bordering the MCR 

(88.77–90.20 Mb) of our finding on chromosome 

9q22. Furthermore, this region corresponds with 

linkage peaks from two previous scans in affected 

sib-pairs. 12,13 With four independent linkage findings 

pointing to the same interval, the existence of a 

relevant gene variation within this region may be 

assumed. Finally, we detected linkage to 16q24.1 

(MCR 82.85–85.12 Mb) in the vicinity of a linkage 

peak also identified by Asherson and colleagues 39 at 

chromosome 16q (LOD > 3.0). Although not referred 

to in detail in the original publications, the finding is 

compatible with previously detected linkage peaks in 

two distinct samples with LOD scores < 2.0. 13,38,40 

Like the interval at chromosome 9q22, the region at 

16q24.1 is highly probable to harbor a susceptibility 

gene for ADHD. 

Additional loci were revealed in individual 

families at 1q25.1, 1q25.3, 12p13.33, 15q11.2-13.3, 

16p12.3-12.2 and 18q11.2-12.3 by parametric and 

non-parametric analyses (Table 3). The locus at 

16p12.3-12.2 (P1) is in close proximity to previously 

reported regions. 9,13,37 Furthermore, the high linkage 

peak at 12p13.33 (P2), which had a relatively high 

MOD score across all families, as well (MODsingle 

2.92, broad phenotype), is close to a known susceptibility 

locus. 37 The two linkage regions at 1q25.1 (P2) 

and 1q25.3 (P8) were also observed by Fisher and 

co-workers. 37 In family P2, two further loci were 

detected, one novel locus at 18p11.2-12.3, the other 

locus at 15q13.1-13.3 replicated twice at that time. 38,41 

The identification of several novel linkage regions 

as well as replication of previously reported loci 

provides further evidence for the highly heterogeneous 

genetic etiology of ADHD. The novel locus at

5q13.1 was detected in an analysis across all families 

and also scored high in two individual families 

suggesting that the identified region contains a 

common gene variant. Approximately 350 kb 

upstream of the linkage peak the PIK3R1 gene is 

located, coding for a regulatory subunit of the 

phosphatidylinositol 3-kinase protein involved in 

post-receptor signaling. PIK3R1 is strongly expressed 

in the prefrontal cortex, a brain region implicated in 

ADHD symptoms and stimulant treatment response. 42 

Although genes within replicated linkage regions 

seem especially promising for fine mapping and 



Figure 2 Genome-wide parametric linkage results in eight families under the narrow phenotype assumption (red, 

MODglobal; blue, MODsingle). 

Figure 3 Genome-wide parametric linkage results in eight families under the broad phenotype assumption (red, 

MODglobal; blue, MODsingle). 

subsequent positional cloning efforts, candidate 

genes may also be found in novel loci, such as the 

syntaxin-binding protein 6 gene (STXBP6) located 

within the MCR of the locus at 14q12. STXBP6 is 

involved in neurotransmitter release, and interacts 

with synaptosome-associated protein receptors 

including SNAP 25. 10 The genes encoding the 

g-aminobutyric-acid (GABA) receptor subunit b-3 

(GABRB3) and the amyloid b-A4 precursor proteinbinding 

member 2 (APBA2) are located in the center 

of the MCR at 15q12-13.1 (P2). GABRB3 is encoding 

a subunit of the GABA/benzodiazepine receptor 

527 


528 


Table 3 Linkage results in individual families 

Locus Info Narrow Broad MCR DI fam n/r 

MODglobal MODsingle NPL MODglobal MODsingle NPL Marker Physical position (in Mb) 



1q25.1 0.95 4.17 4.17 7.34 1.79 1.79 4.33 rs1234315 169.91 1.00 P2 r 

1q25.3 0.92 2.99 2.99 6.00 0.85 0.91 0.68 rs2274984-rs2148620 179.84–180.43 0.98 P8 r 

9q31.1-33.1 0.88 1.50 1.50 3.20 2.59 2.59 9.51 rs1809921-rs7861999 104.89–112.20 1.00 P1 r 

9q33 0.73 2.52 2.52 5.67 1.91 1.91 3.41 rs230150-rs10513454 119.03–125.75 0.98 P5 r 

12p13.33 0.96 3.11 3.11 5.12 2.32 2.32 5.28 rs4980804-rs1009281 0.13–2.40 0.08 P2 r 

15q11.2-13.3 0.95 3.68 3.68 4.68 1.97 1.97 1.84 rs4310812-rs3850097 19.85–31.20 0.68 P2 r 

16p12.3-12.2 0.89 1.41 2.11 1.69 3.84 3.84 4.47 rs10492784-rs226042 19.27–21.20 1.00 P1 r 

18q11.2-12.3 0.95 2.90 2.90 5.31 1.82 1.82 4.09 rs1623173-rs2543029 20.08–41.35 0.34 P2 n 

For further annotations, please see Table 2. 

complex mediating the effects of the major inhibitory 

neurotransmitter, and the APBA2 product is involved 

in synaptic vesicle docking/fusion by interacting with 

STXBP1. 43 

Linkage findings across all eight families resulted 

from the contributions of 3–7 families (family score 

> 0). Particular for the linkage region at 5q13.1 

(MODglobal 4.16, all families, broad phenotype), 

families P1 and P3 also showed relatively high scores 

(MODglobal > 2.7, broad phenotype) under a similar 

genetic model. The other linkage study that investigated 

extended pedigrees segregating for ADHD 14 

only detected a single linkage region at 4q13.2 that 

corresponded in two distinct families. Even though a 

large set of families originating from the same genetic 

isolate (n = 16) was investigated, no further linkage 

peaks across families were found. In our study, all loci 

identified in individual families were replications of 

previous linkage scans in ADHD with the exception of 

the locus at 18p11.2-12.3. Therefore, candidate genes 

at these loci not only may be specific risk factors for 

ADHD in the respective family, but are also likely to 

represent common predisposing factors. The novel 

locus on chromosome 18, however, requires replication. 

Notably, our study replicated results from 

genome-wide scans in affected sib-pairs further 

emphasizing the epidemiological relevance of our 

findings. Overlap with loci identified in the genetic 

isolate was only found for regions obtained by 

analysis of comorbid ODD or CD. 14 Since comorbid 

behavioral problems are frequent in ADHD, comorbidity 

with CD has been suspected to indicate a 

genetically more pronounced variant of ADHD. 35,44 

Pleiotropy and variable genetic dispositions between 

different population groups may account for the 

discrepant results of the two extended pedigree-based 

scans. 14,35 

With the application of a genome-wide highdensity 

scan, our results appear to provide a more 

accurate estimation of localization of candidate gene 

loci than the classical micro-satellite approach. 

Therefore, variance in the location of linkage peaks 

might, in part, be due to these alternative approaches. 

Differences between studies may also result from low 

statistical power to detect true linkage due to small 

sample size. 36 Although the total number of the 

investigated individuals is comparatively small in 

our study, the specific family structure and the dense 

segregation of ADHD substantially increased statistical 

power, and might have facilitated replication of 

previously reported linkage. The pattern of bilineality 

observed in some of our families possibly decreased 

power to detect true linkage, but on the other hand 

reflected an inherent tendency toward assortative 

mating commonly observed for ADHD. Thus, 

including bilineal families might support the notion 

that observed linkage—especially when replicated— 

will be of relevance for the prevalence of ADHD in the 

general population. The loss of power due to genetic 

heterogeneity caused by bilineality can be partially 

counteracted by taking all results of appropriate

statistical methods into consideration. This approach 

compensates for the disadvantage of both the 

parametric analyses assuming a disease-model 

maximizing the LOD score and the non-parametric 

analyses without any model specification, for example, 

possibly overestimating versus possibly underestimating 

linkage in the respective analyses. Overlapping 

results obtained by both parametric (MODglobal and 

MODsingle) and non-parametric (NPL) statistical 

methods are, therefore, supportive of robust findings. 

The linkage scores identified in individual families by 

parametric analysis were comparable to NPL scores. 

Across all families, however, the total NPL scores were 

calculated without considering different degrees of 

informativeness between the families, and, therefore, 

did not match the respective MOD scores. Although, in 

contrast to nearly all previous genome-wide linkage 

scans, we discuss only those findings that met the 

criteria for genome-wide significance suggested by 

Lander and Kruglyak, 34 it ought to be kept in mind that 

all findings are nominal and not adjusted for multiple 

testing. However, since our intention to obtain comparable 

results from various statistical approaches 

accounts for the dilemma of multiple testing, 

Bonferroni correction would be too conservative. 

In conclusion, we detected novel and replicated 

several previously reported linkage loci for ADHD. 

The considerable overlap with earlier genome-wide 

scans indicates that even though the genetic etiology 

of ADHD is complex, there is increasing evidence for 

common gene effects throughout different populations. 

We propose new susceptibility genes with 

possible functional relevance in light of existing brain 

metabolic models of ADHD. This genome-wide 

linkage scan for ADHD employed high marker density 

to optimize resolution and novel strategies of data 

analysis were applied to meet the statistical demands 

imposed by sample structure and marker density. In 

contrast to a previously conducted scan in extended 

pedigrees, 14 the families are unrelated, ascertained on 

the basis of tertiary referral and not part of a genetic 

isolate. The investigation of multigenerational pedigrees 

is one of the most promising approaches to 

identify genetic variants associated with disorders of 

multifactorial etiology, as the phenotype within 

families is rather homogeneous, and erroneous investigation 

of phenocopies is, therefore, highly unlikely. 

While several whole genome association 

(WGA) studies currently underway will likely clarify 

whether common variants explain any of the reported 

linkage peaks, some of our findings may represent the 

effect of rare alleles that are not detected easily by a 

WGA approach. Although specific limitations of the 

study design need to be considered, our results 

encourage ongoing efforts in the investigation of 

multigenerational families with dense segregation of 

genetically complex psychiatric disorders. Future 

investigations will have to include fine mapping, 

positional cloning efforts, WGA studies and metaanalytic 

approaches to clarify the relevance of the 

present findings. 



Acknowledgments 

We thank all families for their participation and 

support. We also greatly appreciate the support from 

several co-workers, who contributed to organization 

of the study, data management and technical assistance: 

Andrea Boreatti-Hümmer, Annette Nowak, 

Gabriela Ortega, Ulrike Schülter, Nicole Steigerwald, 

Theresia Töpner. We thank Professor Konstantin 

Strauch for supporting us in using the program 

GENEHUNTER-MODSCORE. This study was supported 

by the Deutsche Forschungsgemeinschaft 

(DFG: KFO 125, SFB 581, ME 1923/5-1, ME 1923/ 

5-3, GRK 1389/1) and the Bundesministerium für 

Bildung und Forschung (BMBF: 01GV0605). 

Conflict of Interest 

The authors declare no competing financial interests. 

Author contributions: The study was supervised and 

directed by KPL, JM, CF and AW. MR, HP, CS and CJ 

ascertained and clinically characterized the families. 

JR, TR, MH and SW contributed to clinical characterization 

of family members. DWC, RB and DAS carried 

out the genetic analysis. Fine mapping was carried 

out by JS, CV and JM. Data analyses and genotyping 

were performed by KPL, MR, TR, AR and co-workers. 

Statistical analysis was performed by TN, AD and HS; 

CWK contributed to the power analysis. MR, CF, CJ 

and KPL wrote and revised the manuscript. 

References 

1 Faraone SV, Sergeant J, Gillberg C, Biederman J. The worldwide 

prevalence of ADHD: is it an American condition? World 

Psychiatry 2003; 2: 104–113. 

2 Kessler RC, Adler L, Barkley R, Biederman J, Conners CK, 

Demler O et al. The prevalence and correlates of adult ADHD in 

the United States: results from the National Comorbidity Survey 

Replication. Am J Psychiatry 2006; 163: 716–723. 

3 Biederman J, Monuteaux MC, Mick E, Spencer T, Wilens TE, 

Silva JM et al. Young adult outcome of attention deficit 

hyperactivity disorder: a controlled 10-year follow-up study. 

Psychol Med 2006; 36: 167–179. 

4 Jacob CP, Romanos J, Dempfle A, Heine M, Windemuth-Kieselbach C, 

Kruse A et al. Co-morbidity of adult attention-deficit/hyperactivity 

disorder with focus on personality traits and related disorders in a 

tertiary referral center. Eur Arch Psychiatry Clin Neurosci 2007; 

257: 309–317. 

5 Faraone SV, Perlis RH, Doyle AE, Smoller JW, Goralnick JJ, 

Holmgren MA et al. Molecular genetics of attention-deficit/ 

hyperactivity disorder. Biol Psychiatry 2005; 57: 1313–1323. 

6 Faraone SV, Khan SA. Candidate gene studies of attention-deficit/ 

hyperactivity disorder. J Clin Psychiatry 2006; 67(Suppl 8): 13–20. 

7 Friedel S, Saar K, Sauer S, Dempfle A, Walitza S, Renner T et al. 

Association and linkage of allelic variants of the dopamine 

transporter gene in ADHD. Mol Psychiatry 2007; 12: 923–933. 

8 Bakker SC, van der Meulen EM, Oteman N, Schelleman H, 

Pearson PL, Buitelaar JK et al. DAT1, DRD4, and DRD5 polymorphisms 

are not associated with ADHD in Dutch families. Am 

J Med Genet B Neuropsychiatr Genet 2005; 132: 50–52. 

9 Smalley SL, Bailey JN, Palmer CG, Cantwell DP, McGough JJ, 

Del’Homme MA et al. Evidence that the dopamine D4 receptor is a 

susceptibility gene in attention deficit hyperactivity disorder. Mol 

Psychiatry 1998; 3: 427–430. 

529 


530 



10 Faraone SV, Perlis RH, Doyle AE, Smoller JW, Goralnick JJ, 

Holmgren MA et al. Molecular genetics of attention-deficit/ 

hyperactivity disorder. Biol Psychiatry 2005; 57: 1313–1323. 

11 Walitza S, Renner TJ, Dempfle A, Konrad K, Wewetzer C, 

Halbach A et al. Transmission disequilibrium of polymorphic 

variants in the tryptophan hydroxylase-2 gene in attention-deficit/ 

hyperactivity disorder. Mol Psychiatry 2005; 10: 1126–1132. 

12 Hebebrand J, Dempfle A, Saar K, Thiele H, Herpertz-Dahlmann B, 

Linder M et al. A genome-wide scan for attention-deficit/ 

hyperactivity disorder in 155 German sib-pairs. Mol Psychiatry 

2006; 11: 196–205. 

13 Ogdie MN, Macphie IL, Minassian SL, Yang M, Fisher SE, 

Francks C et al. A genomewide scan for attention-deficit/ 

hyperactivity disorder in an extended sample: suggestive linkage 

on 17p11. Am J Hum Genet 2003; 72: 1268–1279. 

14 Arcos-Burgos M, Castellanos FX, Pineda D, Lopera F, Palacio JD, 

Palacio LG et al. Attention-deficit/hyperactivity disorder in a 

population isolate: linkage to loci at 4q13.2, 5q33.3, 11q22, and 

17p11. Am J Hum Genet 2004; 75: 998–1014. 

15 Faraone SV, Biederman J, Monuteaux MC. Toward guidelines for 

pedigree selection in genetic studies of attention deficit hyperactivity 

disorder. Genet Epidemiol 2000; 18: 1–16. 

16 Kaufman J, Birmaher B, Brent D, Rao U, Flynn C, Moreci P et al. 

Schedule for affective disorders and schizophrenia for school-age 

children-present and lifetime version (K-SADS-PL): initial reliability 

and validity data. J Am Acad Child Adolesc Psychiatry 1997; 

36: 980–988. 

17 Schneider S, Margraf J, Unnewehr S. Kinder-DIPS Diagnostisches 

Interview bei Psychischen Störungen von Kindern. Springer: 

Berlin, 1995. 

18 Achenbach T, Rescorla L. Manual for the Child Behavior Checklist/6–18. 

University of Vermont, Department of Psychiatry: 

Burlington, VT, 2001. 

19 Rosler M, Retz W, Retz-Junginger P, Thome J, Supprian T, Nissen T 

et al. Tools for the diagnosis of attention-deficit/hyperactivity 

disorder in adults. Self-rating behaviour questionnaire and 

diagnostic checklist. Nervenarzt 2004; 75: 888–895. 

20 Döpfner M, Lehmkuhl G. Diagnostik-System für psychische 

Störungen im Kindes- und Jugendalter nach ICD-10/DSM-IV 

(DISYPS-KJ), 2nd corrected and extended edition edn. Hans 

Huber: Bern, 2003. 

21 Sheehan DV, Harnett-Sheehan K, Raj BA. The measurement of 

disability. Int Clin Psychopharmacol 1996; 11(Suppl 3): 89–95. 

22 Stiensmeier-Pelster J, Schurmann M, Duda K. Depressions- 

Inventar fU¨r Kinder und Jugendliche (DIKJ), 2nd revised edn. 

Hogrefe: Göttingen, 2000. 

23 Cattell RB, Wei R, Osterland J. Grundintelligenztest Skala 

1—CFT1. Hogrefe: Goettingen, 1997. 

24 Weiss RH. Grundintelligenztest Skala 2—CFT20. Hogrefe: 

Goettingen, 1998. 

25 Lehrl S, Triebig G, Fischer B. Multiple choice vocabulary test 

MWT as a valid and short test to estimate premorbid intelligence. 

Acta Neurol Scand 1995; 91: 335–345. 

26 Retz-Junginger P, Retz W, Blocher D, Weijers HG, Trott GE, 

Wender PH et al. Wender Utah rating scale. The short-version 

for the assessment of the attention-deficit hyperactivity disorder in 

adults. Nervenarzt 2002; 73: 830–838. 

27 Wender P. Attention-Deficit Hyperactivity Disorder in Adults. 

Oxford Press: New York, 1995. 

28 Fydrich T, Renneberg B, Schmitz B, Wittchen H-U. SKID II. 

Interviewheft. Strukturiertes Klinisches Interview fur DSM-IV. 

Achse II (Persönlichkeitsstörungen). Eine deutschsprachige, 


erweiterte Bearbeitung der amerikanischen Originalversion des 

SCID-II von: First, MB, Spitzer, RL, Gibbon, M, Williams, JBW, 

Benjamin, L (Version 3/96). Hogrefe: Göttingen, 1997. 

29 Wittchen H-U, Wunderlich U, Gruschwitz S, Zaudig M. SKID I. 

Strukturiertes Klinisches Interview fur DSM-IV. Achse I: 

Psychische Storungen. Interviewheft und Beurteilungsheft. Eine 

deutschsprachige, erweiterte Bearbeitung der amerikanischen 

Originalversion des SCID I. Hogrefe: Göttingen, 1997. 

30 Ruschendorf F, Nurnberg P. ALOHOMORA: a tool for linkage 

analysis using 10 K SNP array data. Bioinformatics 2005; 21: 

2123–2125. 

31 Abecasis GR, Cherny SS, Cookson WO, Cardon LR. Merlin-rapid 

analysis of dense genetic maps using sparse gene flow trees. Nat 

Genet 2002; 30: 97–101. 

32 Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES. Parametric and 

nonparametric linkage analysis: a unified multipoint approach. 

Am J Hum Genet 1996; 58: 1347–1363. 

33 Strauch K. Parametric linkage analysis with automatic optimization 

of the disease model parameters. Am J Hum Genet 2003; 

73(Suppl 1): A2624. 

34 Lander E, Kruglyak L. Genetic dissection of complex traits: 

guidelines for interpreting and reporting linkage results. Nat 

Genet 1995; 11: 241–247. 

35 Jain M, Palacio LG, Castellanos FX, Palacio JD, Pineda D, Restrepo MI 

et al. Attention-deficit/hyperactivity disorder and comorbid 

disruptive behavior disorders: evidence of pleiotropy and new 

susceptibility loci. Biol Psychiatry 2007; 61: 1329–1339. 

36 Cordell HJ. Sample size requirements to control for stochastic 

variation in magnitude and location of allele-sharing linkage 

statistics in affected sibling pairs. Ann Hum Genet 2001; 65: 

491–502. 

37 Fisher SE, Francks C, McCracken JT, McGough JJ, Marlow AJ, 

MacPhie IL et al. A genomewide scan for loci involved in 

attention-deficit/hyperactivity disorder. Am J Hum Genet 2002; 

70: 1183–1196. 

38 Bakker SC, van der Meulen EM, Buitelaar JK, Sandkuijl LA, Pauls DL, 

Monsuur AJ et al. A whole-genome scan in 164 Dutch sib pairs 

with attention-deficit/hyperactivity disorder: suggestive evidence 

for linkage on chromosomes 7p and 15q. Am J Hum Genet 2003; 

72: 1251–1260. 

39 Asherson P, Zhou K, Anney R, Franke B, Buitelaar J, Ebstein R 

et al. A high density SNP linkage scan with 142 combined 

subtype ADHD sib pairs identifies linkage regions on chromosomes 

9 and 16. Mol Psychiatry 2008; e-pub ahead of print 8 

January 2008. 

40 Ogdie MN, Bakker SC, Fisher SE, Francks C, Yang MH, Cantor RM 

et al. Pooled genome-wide linkage data on 424 ADHD ASPs 

suggests genetic heterogeneity and a common risk locus at 5p13. 

Mol Psychiatry 2006; 11: 5–8. 

41 Ogdie MN, Fisher SE, Yang M, Ishii J, Francks C, Loo SK et al. 

Attention deficit hyperactivity disorder: fine mapping supports 

linkage to 5p13, 6q12, 16p13, and 17p11. Am J Hum Genet 2004; 

75: 661–668. 

42 Niswender KD, Morton GJ, Stearns WH, Rhodes CJ, Myers Jr MG, 

Schwartz MW. Intracellular signalling. Key enzyme in leptininduced 

anorexia. Nature 2001; 413: 794–795. 

43 Okamoto M, Sudhof TC. Mints, Munc18-interacting proteins 

in synaptic vesicle exocytosis. J Biol Chem 1997; 272: 

31459–31464. 

44 Thapar A, Harrington R, McGuffin P. Examining the comorbidity 

of ADHD-related behaviours and conduct problems using a twin 

study design. Br J Psychiatry 2001; 179: 224–229.

Genome-wide linkage analysis of ADHD using high- density SNP ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?