Abstract

Benign recurrent vertigo (BRV) is a common disorder affecting up to 2% of the adult population and may be etiologically related to migraine because of similarities in the clinical spectrum of the phenotypes and a high co-morbidity within families. Many families have multiple-affected genetically related individuals suggesting familial transmission of the disorder with moderate to high penetrance. While clinically similar to episodic ataxias, there are currently no genes identified that contribute to BRV and no systematic linkage studies performed. In an initial effort to genetically define BRV, we have selected from our Neurology Clinic population a subset of 20 multigenerational families with apparent autosomal dominant transmission, and performed genetic linkage mapping using both parametric and non-parametric linkage (NPL) approaches. The Affymetrix 10K SNP Mapping Assay was used for the genotyping. Heterogeneity LOD (HLOD) analysis reveals the evidence of genetic heterogeneity for BRV and evidence of linkage in a subset of the families to 22q12 (HLOD=4.02). An additional region was identified by NPL analysis at 5p15 (LOD=2.63). As migraine is observed substantially more commonly both within the BRV-affected individuals and the related family members, it is possible that a form of migraine is allelic to the BRV locus at 22q12. However, testing linkage or the chromosome 22q12 region to a broader migraine/vertigo phenotype by defining affectation status as either migrainous headaches or BRV greatly weakened the linkage signal, and no significant other peaks were detected. Thus, BRV and migraine does not appear to be allelic disorders within these families. We conclude that BRV is a heterogeneous genetic disorder, appears genetically distinct from migraine with aura and is linked to 22q12. Additional family and population-based linkage and association studies will be needed to determine the causative alleles.

INTRODUCTION

The population prevalence of recurrent dizziness is probably more than 20% (1), as it can result from diverse causes (2). However, chronic recurrent vertigo is less common and may affect up to 2% of the general population (3). The majority of individuals with chronic recurrent vertigo has no identifiable cause, and has no progression of the disorder, and is thus classified as benign recurrent vertigo (BRV). This disorder, which is not associated with neurological or auditory signs, has several other names in the literature including benign paroxysmal vertigo of childhood, BRV of adulthood and vestibular Meniere syndrome. We have previously noted that BRV commonly occurs in other family members often in an autosomal dominant pattern (3). In this study, we further restricted our definition of BRV to include evidence of familial transmission with at least three first-degree relatives affected. By restricting our definition of BRV to include this familial aspect, we sought to select families with autosomal dominant transmission of a disease allele and thus improve our chances of detecting linkage within multigenerational families with multiple-affected individuals. As none of the families were sufficiently large by themselves, we have performed a joint analysis across all of the families to determine linkage within the aggregate of 20 families collected. We used the Affymetrix GeneChip® Human Mapping 10K (HM10K) assay which allows the simultaneous genotyping of ∼11 000 SNPs across the genome at a mean intermarker distance of 210 kb and mean heterozygosity of 0.38, which provides high information content throughout most of the genome. This approach provides a rapid and powerful tool for discerning linkage in complex diseases and Mendelian disorders as several studies have already shown (49). We present results from the first genome-wide scan by both parametric linkage (model-based) and non-parametric linkage (NPL) (model-free) analyses of 20 families with a total of 148 individuals genotyped of which 77 are affected with BRV. HLOD and NPL analyses support a model of genetic heterogeneity and provide suggestive evidence for linkage to 22q12.

A secondary aim of our study was to investigate the genetic relationship between BRV and migraine. BRV appears to be phenotypically related to migraine, with similar triggers and a high incidence of co-morbidity. Several linkage-mapping studies have been performed across the human genome for the linkage of migraine with aura (MA) and migraine without aura (MO) as phenotypes with linkage to seven regions, but no replications of findings across studies using the same phenotype have been found. From these available studies it is clear that migraine is genetically heterogeneous. The closest replication is a region on chromosome 4q21-23 which has been linked with MO in one study and MA in another study, perhaps indicating that migraine-related disorders should be considered a broad spectrum disorder. Extending this concept further to include recurrent episodes of vertigo, it has been suggested that BRV and migraine disorders may be allelic disorders (3). Thus, we have analyzed the presented whole genome linkage data to assess if there is evidence that BRV and migraine are allelic in our families. Including both BRV and migraine as ‘affected’ in the linkage results (developing a broad phenotype) lessened support of linkage to the regions identified to BRV alone. Alternatively, using MA or MO as the affected phenotype does not provide any suggestive evidence of linkage to any of the previously reported region of linkage to migraine phenotypes (data not shown). Thus, within these 20 families, there is no strong support that MA or MO are linked to similar locations in the genome as BRV.

RESULTS

BRV linkage analysis and family-based association tests

A genome-wide scan was performed on 20 families containing 257 individuals, which included 148 genotyped individuals after the data cleaning process. Among 12 different models tested, the strongest evidence for a BRV susceptibility locus was observed (HLOD=4.02, α=0.7) under model 9, which assumes 70% penetrance and 1% phenocopy rate (Table 1). The α-value of 0.7 means that 70% of the families demonstrate some evidence of linkage to this region. However, we interpret this value with caution because it is well known that α estimates for complex diseases are not accurate although the HLOD calculation can be robust and powerful in detecting linkage under heterogeneity (10,11). Although the highest HLOD score was seen in model 9, model 1 and model 5, which used the same phenocopy rate of 1%, yielded similar HLOD scores (3.3 and 3.9, respectively) at the same region and all three models under phenocopy rate of 5% also generated peaks at the same region (Table 2). A summary of the statistics for model 9 is shown in Figure 1 across the entire genome. The chromosome 22q12 linkage is supported by NPL analysis (NPL=2.37, Fig. 2). A single additional region at 5p15 (NPL=2.53) is indicated from the NPL analysis (Fig. 3). Thus, we detect two peaks above our empirical threshold of 1.86 of observing one false positive signal per genome-scan. Peaks on chromosome 22 and chromosome 5 seem to be independent from each other without any systemic grouping of families that contribute to both of the peaks detected. With higher phenocopy rates under model 11, another linkage peak reached a suggestive linkage threshold: chromosome 3q24 (HLOD=1.91, NPL=1.53). However, no other model generated a HLOD over 1.86.

Under the null hypothesis of ‘no association in the presence of linkage’, single-marker associations with the BRV trait were tested for the markers under the 1-LOD support intervals of all three regions showing some evidence of linkage. Among 17 markers on chromosome 22 and 31 markers on chromosome 5, only SNP_A-1513383 (A/G) allele G showed nominal association (six informative families, Z=2.359, P-value=0.01) which we interpret as non-significant.

Relationship of BRV to migraine

Although International Headache Society (IHS) does not include the diagnosis of migrainous vertigo currently, vertigo is commonly reported by patients with migraine (with or without visual aura). The association of migraine and vertigo, both periodic neurologic conditions, has been reported by several investigators over the last few decades (3,1217). Similarly, in our study, of the 221 family members with known phenotypes, 113 report migraine headaches (48%) and about 50% of them report MA. Only 5/49 (10%) of the married-in individuals within the pedigree report migraine symptoms, which is characteristic of the US population at large. Thus, the rate of migraines in relatives of individuals is substantially higher than reported in the population at large (16%) (18) (one-side proportion test: P-value <0.0001). In addition, 79% of the individuals diagnosed with BRV were also diagnosed with migraine (MA or MO). We previously speculated that the traits of BRV and migraine are genetically related and may be allelic (3). To explore this relationship on the basis of the linkage signals, we used alternate phenotypes to search for linkage signals from across the set of 20 families. For instance, if the linkage signals are stronger, using a ‘broader’ phenotype which includes MA or MO (or both) in addition to BRV, then we would take this as evidence that the disorders may share susceptibility loci. Parametric and NPL analyses were performed by broadening the affected status to include family members with migraines. Using this broader phenotype, the linkage signals on chromosome 5 and 22 were diminished, and no new locus reached a significant level. Similarly, by using only MA as the sole affectation status, neither of the two suggestive linkage peaks with BRV alone was detected above a suggestive level. However, one peak on chromosome 16q12 (NPL=1.05) detected using only BRV was strengthened by including MA as the affectation status (NPL=2.05, data not shown). It is possible that this chromosomal locus may harbor a broader susceptibility gene for migraine and vertigo. However, the chance occurrence of such overlaps is high.

DISCUSSION

We performed the first dense genome-wide linkage scan for BRV susceptibility loci using 20 multiplex pedigrees. The data provide strong evidence for genetic heterogeneity as a subset of the families are excluded from being linked to the reported linkage peak on 22q12. None of the individual families was sufficiently large to generate independent and significant evidence of linkage under an assumption of autosomal dominant transmission. Under a model of substantial genetic heterogeneity, which is readily plausible, each family could have a distinct genetic susceptibility locus with incomplete penetrance. Several regions of interest provide evidence of linkage by HLOD and NPL analyses across a subset of the families, most notably the 22q12 region. Under all models of penetrance tested with low but reasonable phenocopy rates, significant linkage is detected only to the BRV phenotype. However, we interpret this linkage signal with some caution. By testing multiple closely related models and maximizing over unknown heterogeneity using multipoint analysis, our type I error rate was increased. In order to place the results in proper context, Greenberg and Abreu (19) showed that the critical threshold should be increased by 0.7 for multipoint HLOD linkage analysis instead of two-point linkage analysis. Thus, the traditional conservative LOD threshold of 3.3 (20) should be increased to 4.0. Thus, our linkage peak on chromosome 22q12 should be considered suggestive at this point.

BRV is phenotypically related to periodic ataxia syndromes like EA-1 and EA-2, which are autosomal dominant disorders shown to be caused by point mutations in ion channels. We speculate that ion channels or channel-associated proteins (21) within these regions of linkage would be primary candidates for sequence and association analysis. There are two interesting genes within the 1-LOD support interval on chromosome 22:adenosine A2a receptor [ADORA2A (MIM 102776)] and beta-adrenergic receptor kinase 2 (ADRBK2 [MIM 109636]). ADORA2A is abundant in basal ganglia and vasculature, and it stimulates adenylyl cyclase. It is a major target of caffeine and the ADORA2A-deficient mice shows a neuroprotective phenotype (22,23). ADRBK2, also known as G-protein-coupled receptor kinase 3 (GRK3), is expressed throughout the brain, abundantly at the synapses playing a role in desensitization of synaptic receptors (24). Within the chromosome 5p15 region is adenylate cyclase 2 (ADCY2 [MIM*103071]) which is expressed in brain. Secondary messengers like cAMP, catalyzed by adenylate cyclases, represent a putative feedback mechanism to regulate ion fluxes in excitable cells (25) through indirect or direct interactions with channel proteins.

While MO and MA are substantially more prevalent within these 20 families than the general population, there is no evidence that migraine is allelic with BRV within these families. The reported suggestive linkage peak on chromosome 16q12 does not overlap with any known migraine loci, and there was no support (nominal P-value ≤0.01) for any of the previously reported migraine loci for MA or MO (2638). These findings are consistent with the heterogeneous finding reported to date in migraine genetics.

From our data and experience with other common human disorders, it is not clear whether multiple rare variants (like EA-1 and EA-2) or common variants in specific genes are more likely to contribute to the development of BRV. We believe that both approaches should be pursued simultaneously in order to identify the causative alleles. Although on clinical grounds there seems to be a clear relationship between MA and BRV, the chromosome 22 locus for BRV was not enhanced when MA was used to determine affectation status. However, another locus on chromosome 16 was augmented and reached suggestive linkage when linkage to both BRV and MA was tested. Either there are multiple BRV loci, some associated with migraine and others not associated with migraine, or the genetic relationship between migraine and BRV is more complex than a shared allele. We consider this an initial exploration of a genome-wide approach to studying the genetic contributions to recurrent vertigo and its relationship with migraine phenotypes. Additional family collection and population-based linkage and association studies will be needed to determine the causative alleles. Insights gained by identifying genes involved in heterogeneous BRV are likely to impact our knowledge of neuronal homeostasis and vestibular function. Further, because there is substantial phenotypic and familial overlap with migraine, the genetics of BRV are likely to influence our understanding of other periodic neurologic disorders such as migraine.

MATERIALS AND METHODS

Patient material

Individuals seen in the Neurology Clinic were consented to participate in the MIRB approved research protocol. One hundred and nineteen potential participants were diagnosed in the clinic. All probands underwent extensive neurologic evaluation to exclude identifiable causes of recurrent vertigo. None of the probands had hearing loss or other signs of Meniere's disease. None of the probands seen in the clinic had abnormal neurological signs, and all had at least three spontaneous episodes of vertigo lasting more than a minute. Each proband completed a questionnaire relating to their episodes of vertigo and described features of MA. In addition a detailed family history was obtained. Of the 119 patients diagnosed with BRV, 23 identified at least two additional family members with recurrent vertigo. The proband identified family members willing to participate, and each was interviewed by phone using a structured questionnaire (available on request). Families with three or more affected individuals based on reported recurrent episodes of vertigo (at least three lasting >1 min) were selected for inclusion in this study in order to select for a more familial trait. A total of 20 families were identified and agreed to participate (Table 3). The average family size was ∼13 individuals per family ranging from 6 to 20 individuals per family across an average of 3.6 generations (Fig. 4). Among these individuals, 87 subjects were diagnosed with BRV. Of the BRV individuals 69 also met criteria for MA or MO by IHS criteria. In addition, there were 25 individuals with MA and not BRV and 24 individuals with MO and not BRV. Not all individuals were available for genotyping. Thirty-six subjects were classified as unknown because of inadequate information. Similar to migraine populations and as previously reported (16), there was a sex bias with BRV being more prevalent in females in our sample set.

Genomic DNA preparation and HM10K assay

Genomic DNA from each individual was extracted from peripheral blood, with the PureGene kit (Gentra System) and diluted to 50 ng/µl with TE buffer (0.1 mm EDTA, 10 mm Tris–HCl, pH 8.0). The HM10K assay was performed according to Affymetrix GeneChip® Mapping Assay Manual (Affymetrix, Santa Clara, CA) within the UCLA DNA Microarray Facility. The samples were hybridized on one of the three versions of HM10K array, HM10K Xba 131 (v.1.0) array, HM10K Xba 142 Early Access (v.2.0ea) array and HM10K Xba 142 (v.2.0) array. One hundred and forty-eight HM10K assays were performed on available DNA samples from 39 male and 109 female subjects. On an average, seven DNA samples were analyzed per family.

Data cleaning

The average call rate for a total of 148 HM10K assays was ∼95% (SD: 3.86%), and the distribution of genotype calls was approximately equal among AA, AB and BB genotypes. The SNP calls generated by GeneChip® DNA Analysis Software (GDAS) were converted to different formats by custom scripts for each program used (H. Lee, available on request). Because some of the SNPs on earlier versions of the arrays were replaced or removed for the newer versions of the arrays, only the SNPs that were present in all three versions of arrays were selected, resulting in 9948 SNPs across the genome. SNPs were first ordered by physical map (NCBI build 35) position and then checked for deCode genetic map position provided by Affymetrix. SNPs having the same decode genetic map position were repositioned by extrapolating from Marshfield genetic map provided by Affymetrix. The ordering of these three SNP maps was in concordance. Allele frequencies of the SNPs were derived from all the founders and random individuals in our sampling by using RECODE version 1.4 program.

The accuracy of the genotyping method across these remaining SNPs is high, and was estimated in two ways. First, there were no heterozygous calls in 39 male X-chromosomes. This represents a total sample of 9945 appropriate homozygous SNP calls. Second, Mendelian error rate detected by PEDCHECK (39) was 0.05% after removing four subjects with erroneous reported relationships, and MERLIN (Multipoint Engine for Rapid Likelihood Inference) software (40) was used to check for non-Mendelian errors, which are caused by improbable double recombination and excessive crossovers. The SNPs with either Mendelian error or non-Mendelian error were marked as ‘NoCall’ across all the genotyped subjects within a family. Using the SNP data, four erroneous relationships were identified and the inappropriate samples removed from further analysis. The SNPs that have less than 80% call rate in the overall sample were removed (471 SNPs) as well as those which deviate from Hardy–Weinberg equilibrium (105 SNPs, P<0.01) (tested using PEDSTATS with all the genotyped individuals). Finally, SNPs in strong linkage disequilibrium (LD) with adjacent SNPs were also removed (797 SNPs) because most of the currently used genetic softwares assume linkage equilibrium between all markers and are shown to be unreliable when there is strong LD between markers (41). MERLIN software was used to construct haplotypes for 37 founders with the best option. Restricted to SNPs with minor-allele frequencies of at least 5%, HAPLOVIEW (42) was used to compute the pairwise LD measure (r2) between each SNP and all other SNPs within 500 kb distance from it. The threshold for high LD was set to r2>0.7, and one SNP with the highest call rate was selectively included in the analyses (6). Finally, 8575 SNPs were used for the analyses presented.

Linkage analysis

Both parametric and NPL analyses were performed using pre-release version of Merlin 1.0-alpha (43). Data were converted to MERLIN format by custom scripts (Lee H., available on request) so that each chromosome could be analyzed in each run except for chromosome 2, which was broken into two segments to overcome the memory constraints. For parametric analysis, HLOD score and the fraction (α) of the families linked to certain loci as well as the information content at the position of each marker were calculated. Multiple models were tested to find the parameters that resulted in optimal HLOD score with the gene frequency fixed at 2% based on the estimated population frequency of BRV (3). The parametric analysis was run under an autosomal dominant model: The penetrance for homozygous and heterozygous disease-causing gene carriers were set to 90%, 80% or 70%, and the phenocopy rate for non-gene carriers were set to 1% at the lowest to 20% at the highest. This generated 12 different combinations of penetrance and phenocopy rates. For all the models, α-value was allowed to vary to maximize the HLOD score for the linkage analysis and then for one of the models that resulted in giving the highest HLOD score.

Multipoint NPL analysis was run with Sall statistics that tests for allele sharing among all the affected individuals. The NPL LOD score that MERLIN calculates by the use of the Kong and Cox (44) linear model was extracted to generate genome-wide scan graphical plots. To estimate empirical significance level of our data set, we used MERLIN to generate 100 replicates of the entire genome retaining the family structure. The value of threshold for observing a genome-wide type I error one time per scan was 1.86 for NPL LOD scores under the hypothesis of no linkage. This empirical threshold is close to the values suggested by previous reports (20,45).

Family-based association analysis

Family-based association tests (FBAT) software (46) was used to test the association between BRV and the SNPs within the suggestive linkage regions on chromosome 5 and 22. Although FBAT handles pedigrees by breaking them into all possible nuclear families, under the null hypothesis of ‘no association in the presence of linkage’, sibling marker genotypes should be correlated and nuclear families broken from a single pedigree and thus should no longer be treated as independent. This is accomplished in FBAT through an option calculating the empirical correction to the variance. As a dichotomous trait with the offset set to 0, only the affected subjects contributed to the test statistics. Unaffected subjects were included in the pedigree file to help determine the genotype of the offspring when parental genotypes were missing. Considering that more than half of our study sample pedigrees are extended pedigrees with multiple-affected offspring within one nuclear family, the minimum number of informative individuals within each family was set to 5. Markers within the 1-LOD support interval of each peak were tested for 79 nuclear families under the dominant model. A total of 31 markers on chromosome 5p15 and 17 markers on chromosome 22q12 were tested.

ACKNOWLEDGEMENTS

We are grateful to the patients and their families for participating in this research study. This study was supported by a program project grant from the NIH (DC05524 [R.W.B.]). The work was performed with the assistance of the UCLA DNA Microarray Facility, and we thank members of the Nelson Lab for guidance and input.

Conflict of Interest statement. The authors have no conflicts of interest.

ELECTRONIC-DATABASE INFORMATION

URLs for the data presented herein are as follows:

Affymetrix, http://www.affymetrix.com

FBAT, http://www.biostat.harvard.edu/~fbat/default.html

MERLIN Center for Statistical Genetics, http://csg.sph.umich.edu/pn/index.php?furl=/abecasis/Merlin/index.html

NCBI, http://www.ncbi.nlm.nih.gov

Online Mendelian Inheritance in Man, http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM

PEDSTATS, http://www.sph.umich.edu/csg/abecasis/PedStats/index.html

RECODE, http://watson.hgen.pitt.edu/register/docs/recode.html

Stanford SOURCE, http://source.stanford.edu/cgi-bin/source/sourceSearch

UCSC, http://www.genome.ucsc.edu

Figure 1. Results of genome-wide multipoint linkage scan. Multipoint HLOD scores are plotted along the y-axis and chromosomal position is plotted along the x-axis. Chromosomes are listed in numerical order across the top.

Figure 1. Results of genome-wide multipoint linkage scan. Multipoint HLOD scores are plotted along the y-axis and chromosomal position is plotted along the x-axis. Chromosomes are listed in numerical order across the top.

Figure 2. HLOD and NPL LOD score for chromosome 22. HLOD of model 9 and NPL LOD score are plotted along the y-axis and genetic map positions of chromosome 22 are plotted along the x-axis.

Figure 2. HLOD and NPL LOD score for chromosome 22. HLOD of model 9 and NPL LOD score are plotted along the y-axis and genetic map positions of chromosome 22 are plotted along the x-axis.

Figure 3. Results of genome-wide NPL analysis and information analysis. NPL LOD score is plotted along the y-axis and chromosomal position is plotted along the x-axis. Chromosomes are listed in numerical order across the top.

Figure 3. Results of genome-wide NPL analysis and information analysis. NPL LOD score is plotted along the y-axis and chromosomal position is plotted along the x-axis. Chromosomes are listed in numerical order across the top.

Figure 4. Pedigrees of the 20 families studied. Individuals affected with BRV, MO and/or MA are shown. Probands are denoted with a filled arrowhead. Although the pedigree notated with * is shown to have only two affected with BRV, it was included in the analysis because another family member, who was not yet genotyped for this study, has been diagnosed with BRV.

Figure 4. Pedigrees of the 20 families studied. Individuals affected with BRV, MO and/or MA are shown. Probands are denoted with a filled arrowhead. Although the pedigree notated with * is shown to have only two affected with BRV, it was included in the analysis because another family member, who was not yet genotyped for this study, has been diagnosed with BRV.

Table 1.

Summary of BRV genome-wide linkage scan showing HLOD and NPL score

Chromosome Genetic distance (cM) Cytogenic location HLOD (α)a NPLb 
102 3p14 – 1.85 
192 3q24 1.3 (0.35) 1.53 
113 4q21 1.25 (0.18) – 
16 5p15 0.48 2.63c 
10 135 10q23 1.56 1.02 
15 128 15q21 – 0.89 
16 90 16q12 – 1.05 
17 15 17p13 – 0.37 
19 10 19p13 1.53 1.18 
22 32 22q11 4.02 (0.7)c 2.37c 
22 42 22q12 1.64 (0.22) 1.07 
Chromosome Genetic distance (cM) Cytogenic location HLOD (α)a NPLb 
102 3p14 – 1.85 
192 3q24 1.3 (0.35) 1.53 
113 4q21 1.25 (0.18) – 
16 5p15 0.48 2.63c 
10 135 10q23 1.56 1.02 
15 128 15q21 – 0.89 
16 90 16q12 – 1.05 
17 15 17p13 – 0.37 
19 10 19p13 1.53 1.18 
22 32 22q11 4.02 (0.7)c 2.37c 
22 42 22q12 1.64 (0.22) 1.07 

aLinkage peaks over 1.5 from HLOD (model 9) and NPL analyses are shown here.

bNPL LOD score was calculated by the use of the Kong and Cox linear model.

cPeaks showing suggestive evidence of linkage.

Table 2.

Maximum HLOD scores at the 22q12 linkage peak in all 12 models tested

Phenocopy (Penetrance) rate 1% 5% 10% 20% 
90% 3.267 1.672 0.811 0.384 
80% 3.867 2.092 1.127 0.484 
70% 4.024 2.329 1.277 0.555 
Phenocopy (Penetrance) rate 1% 5% 10% 20% 
90% 3.267 1.672 0.811 0.384 
80% 3.867 2.092 1.127 0.484 
70% 4.024 2.329 1.277 0.555 
Table 3.

Summary of BRV pedigree sample

Sample description Count 
Total families  20 
Total individuals 257 
Total nuclear families  79 
Male 108 (42%) 
Female 149 (58%) 
Genotyped individuals 148 (58%) 
Phenotype assigned 221 (86%) 
BRVa  87 (39%)c 
BRV Male:Female 19:68 (1:3.57) 
MO (Migraine without aura)b  59 (27%)c 
MO Male:Female 16:43 (1:2.68) 
MA (Migraine with aura)b  54 (24%)c 
MA Male:Female 11:43 (1:3.91) 
BRV or MO or MA 136 (53%)c 
Sample description Count 
Total families  20 
Total individuals 257 
Total nuclear families  79 
Male 108 (42%) 
Female 149 (58%) 
Genotyped individuals 148 (58%) 
Phenotype assigned 221 (86%) 
BRVa  87 (39%)c 
BRV Male:Female 19:68 (1:3.57) 
MO (Migraine without aura)b  59 (27%)c 
MO Male:Female 16:43 (1:2.68) 
MA (Migraine with aura)b  54 (24%)c 
MA Male:Female 11:43 (1:3.91) 
BRV or MO or MA 136 (53%)c 

aIncluding subjects diagnosed with MO or MA.

bIncluding subjects diagnosed with BRV.

cPercentage of affected individuals from population assigned with BRV or MO or MA.

References

1
Kroenke, K. and Price, R.K. (
1993
) Symptoms in the community. Prevalence, classification, and psychiatric comorbidity.
Arch. Intern. Med.
 ,
153
,
2474
–2480.
2
Baloh, R.W. (
1998
) Vertigo.
Lancet
 ,
352
,
1841
–1846.
3
Oh, A.K., Lee, H., Jen, J.C., Corona, S., Jacobson, K.M. and Baloh, R.W. (
2001
) Familial benign recurrent vertigo.
Am. J. Med. Genet.
 ,
100
,
287
–291.
4
John, S., Shephard, N., Liu, G., Zeggini, E., Cao, M., Chen, W., Vasavda, N., Mills, T., Barton, A., Hinks, A. et al. (
2004
) Whole-genome scan, in a complex disease, using 11 245 single-nucleotide polymorphisms: comparison with microsatellites.
Am. J. Hum. Genet.
 ,
75
,
54
–64.
5
Middleton, F.A., Pato, M.T., Gentile, K.L., Morley, C.P., Zhao, X., Eisener, A.F., Brown, A., Petryshen, T.L., Kirby, A.N., Medeiros, H. et al. (
2004
) Genomewide linkage analysis of bipolar disorder by use of a high-density single-nucleotide-polymorphism (SNP) genotyping assay: a comparison with microsatellite marker assays and finding of significant linkage to chromosome 6q22.
Am. J. Hum. Genet.
 ,
74
,
886
–897.
6
Schaid, D.J., Guenther, J.C., Christensen, G.B., Hebbring, S., Rosenow, C., Hilker, C.A., McDonnell, S.K., Cunningham, J.M., Slager, S.L., Blute, M.L. et al. (
2004
) Comparison of microsatellites versus single-nucleotide polymorphisms in a genome linkage screen for prostate cancer-susceptibility loci.
Am. J. Hum. Genet.
 ,
75
,
948
–965.
7
Lee, H., Wang, H., Jen, J.C., Sabatti, C., Baloh, R.W. and Nelson, S.F. (
2004
) A novel mutation in KCNA1 causes episodic ataxia without myokymia.
Hum. Mutat.
 ,
24
,
536
.
8
Sellick, G.S., Longman, C., Tolmie, J., Newbury-Ecob, R., Geenhalgh, L., Hughes, S., Whiteford, M., Garrett, C. and Houlston, R.S. (
2004
) Genomewide linkage searches for Mendelian disease loci can be efficiently conducted using high-density SNP genotyping arrays.
Nucleic Acids Res.
 ,
32
,
e164
.
9
Uhlenberg, B., Schuelke, M., Ruschendorf, F., Ruf, N., Kaindl, A.M., Henneke, M., Thiele, H., Stoltenburg-Didinger, G., Aksu, F., Topaloglu, H. et al. (
2004
) Mutations in the gene encoding gap junction protein alpha 12 (connexin 46.6) cause Pelizaeus-Merzbacher-like disease.
Am. J. Hum. Genet.
 ,
75
,
251
–260.
10
Whittemore, A.S. and Halpern, J. (
2001
) Problems in the definition, interpretation, and evaluation of genetic heterogeneity.
Am. J. Hum. Genet.
 ,
68
,
457
–465.
11
Hodge, S.E., Vieland, V.J. and Greenberg, D.A. (
2002
) HLODs remain powerful tools for detection of linkage in the presence of genetic heterogeneity.
Am. J. Hum. Genet.
 ,
70
,
556
–559.
12
Neuhauser, H., Leopold, M., von Brevern, M., Arnold, G. and Lempert, T. (
2001
) The interrelations of migraine, vertigo, and migrainous vertigo.
Neurology
 ,
56
,
436
–441.
13
Slater, R. (
1979
) Benign recurrent vertigo.
J. Neurol. Neurosurg. Psychiat.
 ,
42
,
363
–367.
14
Moretti, G., Manzoni, G.C., Caffarra, P. and Parma, M. (
1980
) ‘Benign recurrent vertigo’ and its connection with migraine.
Headache
 ,
20
,
344
–346.
15
Kayan, A. and Hood, J.D. (
1984
) Neuro-otological manifestations of migraine.
Brain
 ,
107
,
1123
–1142.
16
Baloh, R.W. and Jen, J.C. (
2002
) Genetics of familial episodic vertigo and ataxia.
Ann. NY Acad. Sci.
 ,
956
,
338
–345.
17
von Brevern, M., Zeise, D., Neuhauser, H., Clarke, A.H. and Lempert, T. (
2005
) Acute migrainous vertigo: clinical and oculographic findings.
Brain
 ,
128
,
365
–374.
18
Rasmussen, B.K., Jensen, R., Schroll, M. and Olesen, J. (
1991
) Epidemiology of headache in a general population—a prevalence study.
J. Clin. Epidemiol.
 ,
44
,
1147
–1157.
19
Greenberg, D.A. and Abreu, P.C. (
2001
) Determining trait locus position from multipoint analysis: accuracy and power of three different statistics.
Genet. Epidemiol.
 ,
21
,
299
–314.
20
Lander, E. and Kruglyak, L. (
1995
) Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results.
Nat. Genet.
 ,
11
,
241
–247.
21
Kullmann, D.M. (
2002
) The neuronal channelopathies.
Brain
 ,
125
,
1177
–1195.
22
Ledent, C., Vaugeois, J.M., Schiffmann, S.N., Pedrazzini, T., El Yacoubi, M., Vanderhaeghen, J.J., Costentin, J., Heath, J.K., Vassart, G. and Parmentier, M. (
1997
) Aggressiveness, hypoalgesia and high blood pressure in mice lacking the adenosine A2a receptor.
Nature
 ,
388
,
674
–678.
23
Chen, J.F., Huang, Z., Ma, J., Zhu, J., Moratalla, R., Standaert, D., Moskowitz, M.A., Fink, J.S. and Schwarzschild, M.A. (
1999
) A(2A) adenosine receptor deficiency attenuates brain injury induced by transient focal ischemia in mice.
J. Neurosci.
 ,
19
,
9192
–9200.
24
Arriza, J.L., Dawson, T.M., Simerly, R.B., Martin, L.J., Caron, M.G., Snyder, S.H. and Lefkowitz, R.J. (
1992
) The G-protein-coupled receptor kinases beta ARK1 and beta ARK2 are widely distributed at synapses in rat brain.
J. Neurosci.
 ,
12
,
4045
–4055.
25
Gusovsky, F. and Daly, J.W. (
1988
) Formation of second messengers in response to activation of ion channels in excitable cells.
Cell. Mol. Neurobiol.
 ,
8
,
157
–169.
26
Ducros, A., Joutel, A., Vahedi, K., Cecillon, M., Ferreira, A., Bernard, E., Verier, A., Echenne, B., Lopez de Munain, A., Bousser, M.G. et al. (
1997
) Mapping of a second locus for familial hemiplegic migraine to 1q21-q23 and evidence of further heterogeneity.
Ann. Neurol.
 ,
42
,
885
–890.
27
Gardner, K., Barmada, M.M., Ptacek, L.J. and Hoffman, E.P. (
1997
) A new locus for hemiplegic migraine maps to chromosome 1q31.
Neurology
 ,
49
,
1231
–1238.
28
Lea, R.A., Shepherd, A.G., Curtain, R.P., Nyholt, D.R., Quinlan, S., Brimage, P.J. and Griffiths, L.R. (
2002
) A typical migraine susceptibility region localizes to chromosome 1q31.
Neurogenetics
 ,
4
,
17
–22.
29
Vanmolkot, K.R., Kors, E.E., Hottenga, J.J., Terwindt, G.M., Haan, J., Hoefnagels, W.A., Black, D.F., Sandkuijl, L.A., Frants, R.R., Ferrari, M.D. et al. (
2003
) Novel mutations in the Na+, K+-ATPase pump gene ATP1A2 associated with familial hemiplegic migraine and benign familial infantile convulsions.
Ann. Neurol.
 ,
54
,
360
–366.
30
Soragna, D., Vettori, A., Carraro, G., Marchioni, E., Vazza, G., Bellini, S., Tupler, R., Savoldi, F. and Mostacciuolo, M.L. (
2003
) A locus for migraine without aura maps on chromosome 14q21.2-q22.3.
Am. J. Hum. Genet.
 ,
72
,
161
–167.
31
Wessman, M., Kallela, M., Kaunisto, M.A., Marttila, P., Sobel, E., Hartiala, J., Oswell, G., Leal, S.M., Papp, J.C., Hamalainen, E. et al. (
2002
) A susceptibility locus for migraine with aura, on chromosome 4q24.
Am. J. Hum. Genet.
 ,
70
,
652
–662.
32
Jones, K.W., Ehm, M.G., Pericak-Vance, M.A., Haines, J.L., Boyd, P.R. and Peroutka, S.J. (
2001
) Migraine with aura susceptibility locus on chromosome 19p13 is distinct from the familial hemiplegic migraine locus.
Genomics
 ,
78
,
150
–154.
33
Nyholt, D.R., Curtain, R.P. and Griffiths, L.R. (
2000
) Familial typical migraine: significant linkage and localization of a gene to Xq24-28.
Hum. Genet.
 ,
107
,
18
–23.
34
Carlsson, A., Forsgren, L., Nylander, P.O., Hellman, U., Forsman-Semb, K., Holmgren, G., Holmberg, D. and Holmberg, M. (
2002
) Identification of a susceptibility locus for migraine with and without aura on 6p12.2-p21.1.
Neurology
 ,
59
,
1804
–1807.
35
Russo, L., Mariotti, P., Sangiorgi, E., Giordano, T., Ricci, I., Lupi, F., Chiera, R., Guzzetta, F., Neri, G. and Gurrieri, F. (
2005
) A new susceptibility locus for migraine with aura in the 15q11-q13 genomic region containing three GABA-A receptor genes.
Am. J. Hum. Genet.
 ,
76
,
327
–333.
36
Nyholt, D.R., Morley, K.I., Ferreira, M.A., Medland, S.E., Boomsma, D.I., Heath, A.C., Merikangas, K.R., Montgomery, G.W. and Martin, N.G. (
2005
) Genomewide significant linkage to migrainous headache on chromosome 5q21.
Am. J. Hum. Genet.
 ,
77
,
500
–512.
37
Bjornsson, A., Gudmundsson, G., Gudfinnsson, E., Hrafnsdottir, M., Benedikz, J., Skuladottir, S., Kristjansson, K., Frigge, M.L., Kong, A., Stefansson, K. et al. (
2003
) Localization of a gene for migraine without aura to chromosome 4q21.
Am. J. Hum. Genet.
 ,
73
,
986
–993.
38
Cader, Z.M., Noble-Topham, S., Dyment, D.A., Cherny, S.S., Brown, J.D., Rice, G.P. and Ebers, G.C. (
2003
) Significant linkage to migraine with aura on chromosome 11q24.
Hum. Mol. Genet.
 ,
12
,
2511
–2517.
39
O'Connell, J.R. and Weeks, D.E. (
1998
) PedCheck: a program for identification of genotype incompatibilities in linkage analysis.
Am. J. Hum. Genet.
 ,
63
,
259
–266.
40
Abecasis, G.R., Cherny, S.S., Cookson, W.O. and Cardon, L.R. (
2002
) Merlin–rapid analysis of dense genetic maps using sparse gene flow trees.
Nat. Genet.
 ,
30
,
97
–101.
41
Schaid, D.J., McDonnell, S.K., Wang, L., Cunningham, J.M. and Thibodeau, S.N. (
2002
) Caution on pedigree haplotype inference with software that assumes linkage equilibrium.
Am. J. Hum. Genet.
 ,
71
,
992
–995.
42
Barrett, J.C., Fry, B., Maller, J. and Daly, M.J. (
2005
) Haploview: analysis and visualization of LD and haplotype maps.
Bioinformatics
 ,
21
,
263
–265.
43
Abecasis, G.R. and Wigginton, J.E. (
2005
) Handling marker-marker linkage disequilibrium: pedigree analysis with clustered markers.
Am. J. Hum. Genet.
 ,
77
,
754
–767.
44
Kong, A. and Cox, N.J. (
1997
) Allele-sharing models: LOD scores and accurate linkage tests.
Am. J. Hum. Genet.
 ,
61
,
1179
–1188.
45
Kenealy, S.J., Babron, M.C., Bradford, Y., Schnetz-Boutaud, N., Haines, J.L., Rimmler, J.B., Schmidt, S., Pericak-Vance, M.A., Barcellos, L.F., Lincoln, R.R. et al. (
2004
) A second-generation genomic screen for multiple sclerosis.
Am. J. Hum. Genet.
 ,
75
,
1070
–1078.
46
Laird, N.M., Horvath, S. and Xu, X. (
2000
) Implementing a unified approach to family-based tests of association.
Genet. Epidemiol.
 ,
19
(Suppl. 1),
S36
–S42.