Association of GABRG3 With Alcohol Dependence
Danielle M. Dick, Howard J. Edenberg, Xiaoling Xuei, Alison Goate, Sam Kuperman, Marc Schuckit, Raymond Crowe,
Tom L. Smith, Bernice Porjesz, Henri Begleiter, and Tatiana Foroud
Background: Evidence from human, animal, and in vitro cell models suggests that ?-aminobutyric acid
(GABA), the major inhibitory neurotransmitter in the human central nervous system, is involved in many
of the neurochemical pathways that affect alcohol use, abuse, and dependence. Both linkage and associa-
tion to the region on chromosome 15q that contains a cluster of GABAAreceptor genes have previously
been reported, but the role of these genes in alcoholism remains inconclusive.
Methods: We conducted family-based association analyses by using a large sample of multiplex alcoholic
between alcohol dependence and the GABAAreceptor genes clustered on chromosome 15q. Multiple
single-nucleotide polymorphisms were tested in each of the three chromosome 15q GABAAreceptor
genes: GABRA5, GABRB3, and GABRG3.
Results: Using both classic trio-based analyses and extended-family analyses, we found consistent evi-
dence of association between alcohol dependence and GABRG3. Nearly all single-nucleotide polymor-
phisms across the gene yielded evidence of association, and haplotype analyses were highly significant. No
consistent evidence of association was observed with either GABRA5 or GABRB3, nor was there evidence
for parent-of-origin effects with any of the genes.
Conclusions: These analyses suggest that GABRG3 may be involved in the risk for alcohol dependence.
These findings support the theory that the predisposition to alcoholism may be inherited as a general state
of central nervous system disinhibition/hyperexcitability that results from an altered responsiveness to
Key Words: GABA, Alcohol Dependence, Genetic Analysis, COGA.
the major inhibitory neurotransmitter in the human central
nervous system, is involved in many of the neurochemical
pathways that affect alcohol use and related disorders.
GABA is involved in several of the behavioral effects of
alcohol, including motor incoordination, anxiolysis, seda-
tion, withdrawal signs, and ethanol preference (Buck, 1996;
Grobin et al., 1998). GABAAreceptor agonists tend to
potentiate the behavioral effects of alcohol, whereas
GABAA receptor antagonists attenuate these effects.
GABAAreceptors have also been implicated in ethanol
tolerance and dependence (Grobin et al., 1998). It is not
VIDENCE FROM ANIMAL, human, and in vitro cell
models suggests that ?-aminobutyric acid (GABA),
clear exactly how GABA reception is involved in these
actions of ethanol (Grobin et al., 1998).
Most of the GABAAreceptor genes are organized into
clusters. Chromosome 4p contains the genes GABRA2,
GABRA4, GABRB1, and GABRG1; chromosome 5q con-
tains GABRA1, GABRA6, GABRB2, and GABRG2; and
GABRG3 (LocusLink; National Center for Biotechnology
Information). This study explored the relationship between
the chromosome 15 GABAAreceptor gene cluster and
alcohol dependence. This cluster has not been studied as
extensively as the clusters on chromosome 4 (Edenberg et
al., 2002; Parsian and Zhang, 1999; Porjesz et al., 2002) and
chromosome 5 [reviewed in Dick and Foroud (2003)], al-
though evidence is accumulating to suggest that the
GABAAreceptor genes on chromosome 15 may be in-
volved in alcohol dependence and related phenotypes. Mi-
crosatellite markers in GABRA5 and GABRB3 were previ-
ously tested for association by using parent-offspring trios
selected from the Collaborative Study on the Genetics of
Alcoholism (COGA) sample (Song et al., 2003). Modest
evidence of association between GABRA5 and alcohol de-
pendence, as defined by the International Classification of
Diseases, 10th revision, was found when the sample was
limited to Caucasians. On the basis of evidence from a
GABRG3 were expressed only from the paternal alleles in
From the Indiana University School of Medicine, Indianapolis, Indiana
(DMD, HJE, XX, TF); Washington University, St. Louis, Missouri (AG);
University of Iowa, Iowa City, Iowa (SK, RC); University of California, San
Diego, California (MS, TLS); and State University of New York, Brooklyn,
New York (BP, HB).
Received for publication June 12, 2003; accepted October 1, 2003.
This national collaborative study is supported by NIH Grant U10AA08403
from the NIAAA. Preparation of this manuscript was also supported by
AA13358 (D. Dick) and K02-AA00285 (T. Foroud).
Reprint requests: Tatiana Foroud, PhD, Department of Medical and
Molecular Genetics, Indiana University School of Medicine, 975 W. Walnut
St., IB-130, Indianapolis, IN 46202-5251; Fax: 317-274-3287; E-mail:
Copyright © 2004 by the Research Society on Alcoholism.
ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH
Vol. 28, No. 1
Alcohol Clin Exp Res, Vol 28, No 1, 2004: pp 4–9
hybrid mouse A9 cells containing a single human chromo-
some 15 (Meguro et al., 1997), Edenberg et al. tested for
paternal transmission of GABRA5 and GABRB3 in the
COGA sample and found evidence of association of both
genes with alcohol dependence (Song et al., 2003). In a
case-control study of Caucasian alcoholics and controls, an
association was also reported between GABRB3 and severe
alcoholism, as defined by documented alcohol-induced
bodily damage such as cirrhosis; furthermore, there was a
significant, progressive decrease in the prevalence of the
most frequent allele of GABRB3 in nonalcoholics, less
severe alcoholics, and severe alcoholics (Noble et al., 1998).
The ?2component of the electroencephalogram (EEG), an
endophenotype for alcoholism, also has been linked to the
chromosome 15 region that contains the GABAAreceptor
gene cluster in the sample collected by COGA (Ghosh et
Here we report family-based association analyses using a
large sample of multiplex alcoholic families collected as
part of the COGA study. Our study has several strengths
that build on the existent literature. We used a large family-
based association design, which avoids potential problems
with population stratification that are introduced by case-
control studies. We analyzed both extended families and
classic transmission disequilibrium test (TDT) trios (com-
posed of an affected offspring and his or her parents) to
examine consistency across analytic methods. We tested
multiple single-nucleotide polymorphisms (SNPs) in each
gene and looked for consistent trends across SNPs within
the gene. Many previous studies have tested only one or
two genetic variants in the gene under study; this can lead
to a false-negative finding if that particular marker is not in
linkage disequilibrium (LD) with the variation or variations
that increase risk for disease. Alternatively, a positive as-
sociation could result if that particular marker is in high LD
with variations in a nearby gene; in such a case, although
the signal would be real, the actual gene involved would not
be the one in which the marker resides. These potentially
incorrect conclusions could result because patterns of LD
are not a simple function of physical distance across the
genome (Abecasis et al., 2001; Gabriel et al., 2002). There-
fore, we analyzed LD between SNPs and used this infor-
mation to help interpret the results of association analyses.
By designing the study in this manner and taking these
steps to ensure data quality and consistency, we aimed to
provide more conclusive evidence regarding the role of the
GABAAreceptor genes on chromosomes 15 in affecting
the risk for alcohol dependence.
COGA is a multisite project for which families were collected at six
centers across the United States: Indiana University, State University of
New York Health Science Center, University of Connecticut, University of
Iowa, University of California–San Diego, and Washington University, St.
Louis. Probands identified through inpatient or outpatient alcohol treat-
ment programs at these six sites were invited to participate if they had a
sufficiently large family (usually more than three siblings with parents
available) with two or more members in a COGA catchment area (Reich,
1996). The institutional review boards of all participating institutions
approved the study. A total of 1227 families of alcohol-dependent pro-
bands were recruited for the first stage of the study. Additionally, a sample
of control families, obtained through random sources such as driver’s
license registries and dental clinics, was assessed. These families consisted
of two parents and at least three children over the age of 14 years. All
individuals were administered the Semi-Structured Assessment for the
Genetics of Alcoholism interview (Bucholz et al., 1994; Hesselbrock et al.,
1999). For this study, individuals were diagnosed with alcohol dependence
by using DSM-IV criteria (American Psychiatric Association, 1994).
Multiplex alcoholic families that were not bilineal and had at least two
affected first-degree relatives in addition to the proband were invited to
participate in the more intensive stage of the study, which included
obtaining blood for genetic analyses. Second- and third-degree relatives in
the families were assessed when they were considered to be informative
for the genetic linkage studies. A total of 987 adult individuals from 105
extended families were included in the initial genotyped dataset (Reich et
al., 1998). A replication sample was ascertained and genotyped by follow-
ing identical procedures; it consisted of 1295 individuals from 157 ex-
tended families (Foroud et al., 2000). Thus, 2282 individuals from 262
multiplex alcoholic families were available for genetic analyses. An addi-
tional 1254 individuals from 227 control families also had blood collected
and were included to increase the power of analyses of LD between
SNPs were chosen across each candidate gene from public databases;
we did not restrict ourselves to coding regions or exons, because allele
frequencies for such SNPs are often low. Locations were in most cases
determined from the annotations in the National Center for Biotechnol-
ogy Information human genome assembly. In some cases, position was
determined by BLASTing the sequence against the human genome as-
sembly. Allele frequencies are not usually available for public SNPs, so
SNPs were genotyped on 2 sets of approximately 40 unrelated individuals
from the Coriell Caucasian American and African American samples to
determine approximate allele frequencies; we preferentially chose SNPs
with high heterozygosities. Genotyping was performed by using a modified
single-nucleotide extension reaction, with allele detection by mass spec-
trometry (Sequenom MassArray system, Sequenom, San Diego, CA). A
total of 3536 individuals from the COGA families were genotyped. All
genotypic data were checked for Mendelian inheritance of marker alleles
with the USERM13 (Boehnke, 1991) option of the MENDEL linkage
computer programs; this was then used to estimate marker allele frequen-
cies. Twenty-one SNPs were genotyped in the chromosome 15 GABAA
receptor gene cluster: 6 SNPs in GABRB3, 4 SNPs in GABRA5, and 11
SNPs in GABRG3. The average heterozygosity across the SNPs in chro-
mosome 15 was 0.44. These SNPs, the gene in which they are located, and
their chromosomal positions are shown in Table 1.
Multiplex alcoholic families were used in tests of association between
each of the SNPs and alcohol dependence. The Pedigree Disequilibrium
Test (PDT) (Martin et al., 2000) was used to analyze associations in the
extended pedigrees. The PDT uses data from all available trios in a family,
as well as discordant sibships. It produces two statistics: the PDT-avg,
which averages the association statistic across all families, and the PDT-
sum, which gives greater weight to families with more informative trios
and discordant sibships (Martin et al., 2001). Because the COGA sample
consists of several very large families that might overly influence the
PDT-sum statistic, we report the values from the PDT-avg statistic. We
also conducted classic TDT trio-based analyses by selecting one trio from
each COGA family. We used the program TRANSMIT for these analyses
(Clayton, 1999). Finally, because the existent literature on chromosome 15
GABRG3 ASSOCIATED WITH ALCOHOLISM
GABAAreceptor genes suggested that there may be parental transmission
effects, we also used the program TDTEX, which performs TDT analyses
that allow for the testing of parent-of-origin effects (Statistical Analysis for
Genetic Epidemiology, version 4.0, Department of Epidemiology and
Biostatistics, Rammelkamp Center for Education and Research, Metro-
Health Campus, Case Western University, Cleveland, OH).
Haplotype analyses were conducted, using both the PDT and TRANS-
MIT, across genes that showed evidence of association at the individual
SNP level. We used a sliding-window approach in which groups of three
SNPs were sequentially tested across the gene. Thus, the first analysis tests
the haplotype composed of SNPs 1, 2, and 3; the second haplotype analysis
tests SNPs 2, 3, and 4, and so on. We used haplotypes computed using
Simwalk2 (Sobel and Lange, 1996) as input into the PDT. Families with
more than one individual with a double recombinant event were omitted
from the analyses. More than 90% of all families were retained for all
analyses. TRANSMIT computes haplotypes internally by averaging over
all possible configurations of parental haplotypes and transmissions con-
sistent with the observed data. Possible haplotype assignments are
weighted according to the probability of each assignment by using the
expectation-maximization algorithm. Because several rare haplotypes
were observed and these can bias the ?2statistic yielded by TRANSMIT,
only haplotypes observed at a frequency of at least 10% were used in the
LD between markers was evaluated by using the program GOLD
(Abecasis and Cookson, 2000) with both multiplex alcoholic family data
and control family data. This program uses haplotype input from Sim-
walk2 (Sobel and Lange, 1996) and produces pairwise disequilibrium
measures for all markers entered into the analysis. Families that had more
than one individual with a double recombination event within a gene were
dropped from subsequent analyses. Several measures of LD are produced
by GOLD. We show the degree of LD between markers, as measured by
?2, for regions for which LD was used to help interpret the association
results. The conventional definition for ?2(Devlin and Risch, 1995) is
used by GOLD:
p1? p2? p1? p2?
The p values were not corrected for multiple testing because the SNPs
are not independent. Thus, we conservatively interpreted our results by
requiring consistency across analytic methods and requiring consistency
between the pattern of association results and the pattern of LD across the
The SNPs in GABRG3 provided consistent evidence of
association with alcohol dependence (Table 1). By using
the program TRANSMIT, 7 of the 11 SNPs tested in
GABRG3 provided significant evidence of association (p ?
0.05). Three of the four remaining SNPs showed evidence
of association at p ? 0.10. Although the results of the PDT
did not reach statistical significance, they were generally
consistent with the results from TRANSMIT, with 8 of the
11 SNPs yielding p values ?0.13. LD analyses demon-
strated that the 1 SNP that showed no evidence of associ-
ation with either TRANSMIT or the PDT—rs2288694—
was not in LD with the other SNPs in GABRG3 (Table 2);
all other SNPs were in reasonably high LD. rs2288694 maps
further from the other GABRG3 SNPs, and there are ap-
parent problems with this region of the genome (see the
note to Table 1). Table 3 lists the number of observed
versus expected transmissions for the SNPs in GABRG3, so
the reader can evaluate the degree of deviation from ex-
The results from the sliding-window haplotype analyses
with both the PDT and TRANSMIT yielded further evi-
dence of association (Table 4). The overall ?2statistics
showed trends toward significance at p ? 0.10 for six of the
nine tests performed with the PDT and seven of nine tests
Table 1. SNPs in the Chromosome 15 GABAAReceptor Gene Cluster and Their Association With Alcoholism
Marker Gene PositionHeterozygosityPDT-avg p ValueTRANSMIT p ValueNo. triosTdtex-maternalTdtex-paternal
Statistics are shown for both the PDT, TRANSMIT, and parent-of-origin effects using TDTEX.
Position is from dbSNP114, May 2003.
aThe accuracy of this SNP position is uncertain. There is a gap and an unfinished sequence in this region. rs2288694 maps far from the other GABRG3 SNPs but
is annotated as being within 2 kilobases of the transcript; BLASTing shows it within exon 1 of the messenger RNA. The REFSEQ database shows 10 exons for GABRG3,
but BLASTing the genome (build 33) gives only 9 hits: 1 middle exon is missing. Thus, the exact distance between SNPs is unknown.
* p ? 0.05; p values were not corrected for multiple tests.
DICK ET AL.
with TRANSMIT. Furthermore, the tests for the specific
risk haplotypes, composed of the alleles that were over-
transmitted in the individual SNP analyses, were significant
for nearly all haplotypes tested across both methods of
There were no consistent patterns of association across
any of the other SNPs tested in the two other GABAA
receptor genes on chromosome 15 (Table 1). The results
from the parental origin tests performed in S.A.G.E. sug-
gest that there are no consistent parent-of-origin effects
(Table 1); we believe that the two SNPs that were signifi-
cant in parental origin tests are likely false positives, be-
cause the other SNPs that they are in LD with show no
evidence of parent-of-origin effects.
We have tested for association between the GABAA
receptor genes clustered on chromosome 15 and alcohol
dependence by using a large sample of multiplex alcoholic
families. Our strategy used multiple analytic methods for
family-based designs, tested multiple SNPs in each gene,
and used patterns of LD among the SNPs to interpret
association results and reduce the possibility of false posi-
tives. We found consistent evidence that GABRG3 on chro-
mosome 15 is associated with alcohol dependence. Nearly
all SNPs tested showed evidence of association, at least at
the trend level, with both classic TDT analyses and the
Haplotype analyses were highly significant with both the
TDT and PDT tests. We are encouraged by the consistency
between the PDT and TRANSMIT haplotype analyses
because the haplotypes were computed by using different
methods in each program (Simwalk2 uses Markov chain
Monte Carlo and simulated annealing algorithms versus
the expectation-maximization algorithm in TRANSMIT).
However, the haplotype analyses were not helpful in fur-
ther narrowing the region of GABRG3 that may contain the
putative risk variant. Haplotype analyses showed associa-
tion across the gene, with no particular patterns of greater
significance in any one region.
Greater statistical significance was obtained from the
TDT, which uses only one independent trio from each
family, than with the PDT, which includes information
from additional family members. We speculate that this
may be due to the inclusion of genotypic data from unaf-
fected individuals in the PDT. In the event of association, it
Table 2. LD Between SNPs Within the Chromosome 15 GABAAReceptor Gene GABRG3 as Measured by ?2
Variablers1871019rs3101639 rs3097493rs3101637rs3101636rs140678 rs140679rs2303879 rs3097490 rs3097489
SNPs are shown in order across the gene, corresponding to the chromosomal positions listed in Table 1.
Table 3. Number of Observed vs. Expected Transmissions for Each Allele
Yielded by TRANSMIT for SNPs Across GABRG3
p values are also shown to evaluate the degree of transmission deviation for
each SNP; these correspond to the values listed in Table 1.
Table 4. p Values From Haplotype Analyses With the PDT and TRANSMIT for
All SNPs Tested in GABRG3
OverallHigh risk OverallHigh risk
Haplotype analyses were conducted with a sliding-window approach for
groups of adjacent SNPs. Each p value represents the test of the haplotype
composed of the three SNPs listed before that value. For both the PDT and
TRANSMIT, the first column lists the p value yielded by the overall ?2statistic
testing all haplotypes. The second column represents the p value yielded for the
test of the specific risk haplotype.
GABRG3 ASSOCIATED WITH ALCOHOLISM
would be expected that the overtransmitted allele would be
more prevalent in affected individuals as compared with
unaffected siblings. However, this was not always the case,
which had the effect of decreasing the magnitude of the
PDT statistic. Analyses with discordant sibling pairs should
be interpreted with caution when alcohol dependence is
studied, because no allowance is made for the possibility
that unaffected siblings may not truly be unaffected, but
simply may not have manifested sufficient symptoms to
meet diagnostic criteria. The PDT does not take into ac-
count an unaffected individual’s age or whether he or she
has passed through the period of highest risk.
No consistent evidence of association was seen with ei-
ther of the other GABAAreceptor genes on chromosome
15. The previous studies (Noble et al., 1998; Song et al.,
2003) that found evidence of association with GABRB3 and
GABRA5 both tested microsatellite markers in the genes.
Thus, it is possible that those positive association results
could have resulted from LD with a nearby GABAArecep-
tor gene. It is also puzzling that we found evidence of
paternal transmission in this gene cluster previously but
found no evidence of parent-of-origin effects here. We
have no biological explanation for this. However, our more
detailed analyses of multiple SNPs in all three genes across
the region provide more power to test for such effects than
our previous analyses, which relied on single microsatellite
markers tested in each gene (Song et al., 2003). Therefore,
we were able to more definitively test for paternal trans-
mission in this dataset, and we found no evidence of such
Our results suggest that GABRG3 is the gene most likely
to be responsible for the association observed in our data-
set and that it is likely that variants in this gene contribute
to the risk for alcoholism. The function of the GABRG3
receptor gene is not clearly understood. Therefore, it is
difficult to speculate why this particular gene, located
within a cluster of GABAAreceptor genes, would be asso-
ciated with alcohol dependence. Acute exposure to ethanol
in vitro increases the activity of GABAAreceptor–coupled
chloride channels, which are heteropentamers composed of
several types of subunits (Buck, 1996). In studies of mice,
chronic alcohol consumption caused the messenger RNA
content of Gabrg3 to increase dramatically and similarly in
both withdrawal seizure–prone mice and withdrawal seizur-
e–resistant mice; this led the authors to suggest that ?3may
be involved in the development of tolerance or behavioral
sensitization to ethanol, because these traits do not differ
between strains (Buck, 1996).
Although the mechanism is unclear, the significant asso-
ciation of alcohol dependence with multiple SNPs and
haplotypes in GABRG3 suggests that variations in this gene
are involved in the etiology of alcohol dependence. On the
basis of data from the literature on event-related potentials,
Begleiter and Porjesz (1999) hypothesized that the predis-
position to alcoholism is inherited as a general state of
central nervous system disinhibition/hyperexcitability. This
state of disinhibition/hyperexcitability is associated with a
number of externalizing disorders. They suggested that
alcohol dependence is among these disorders because the
hyperexcitability is alleviated by the use of alcohol, which
provides a normalizing effect. However, the effect is tem-
porary and requires continued use of increasing amounts of
alcohol to achieve this state, putting the individual at higher
risk of developing alcohol problems and dependence.
GABA was proposed to be involved in this pathway be-
cause GABAergic interneurons provide feedback inhibi-
tion to regulate recurrent excitation. Thus, it is plausible
that this state of disinhibition/hyperexcitability could result
from an altered response to GABA. Linkage to EEG in this
chromosome 15 region (Ghosh et al., 2003), taken together
with our finding that alcohol dependence is associated with
genetic variation in GABRG3, provides support for this
COGA (H. Begleiter, State University of New York Health
Science Center at Brooklyn, principal investigator; T. Reich,
Washington University, co-principal investigator; and H. Eden-
berg, Indiana University, co-principal investigator) includes nine
different centers where data collection, analysis, and storage take
place. The nine sites and principal investigators and co-
investigators are as follows: Howard University (R. Taylor); Indi-
ana University (H. Edenberg, J. Nurnberger Jr, P. M. Conneally,
and T. Foroud); Rutgers University (J. Tischfield); Southwest
Foundation (L. Almasy); State University of New York Health
Sciences Center at Brooklyn (B. Porjesz and H. Begleiter); Uni-
versity of California at San Diego (M. Schuckit); University of
Connecticut (V. Hesselbrock); University of Iowa (R. Crowe and
S. Kuperman); and Washington University in St. Louis (T. Reich,
C. R. Cloninger, J. Rice, and A. Goate). Lisa Neuhold serves as
the NIAAA staff collaborator.
Abecasis GR, Cookson WO (2000) GOLD—graphical overview of link-
age disequilibrium. Bioinformatics 16:182–183.
Abecasis GR, Noguchi E, Heinzmann A, Traherne JA, Bhattacharyya S,
Leaves NI, Anderson GG, Zhang Y, Lench NJ, Carey A, Cardon LR,
Moffatt MF, Cookson WO (2001) Extent and distribution of linkage
disequilibrium in three genomic regions. Am J Hum Genet 68:191–197.
American Psychiatric Association (1994) Diagnostic and Statistical Manual
of Mental Disorders: DSM-IV. 4th ed. American Psychiatric Association,
Begleiter H, Porjesz B (1999) What is inherited in the predisposition
toward alcoholism? A proposed model. Alcohol Clin Exp Res 23:1125–
Boehnke M (1991) Allele frequency estimation from pedigree data. Am J
Hum Genet 48:22–25.
Bucholz KK, Cadoret R, Cloninger CR, Dinwiddie SH, Hesselbrock VM,
Nurnberger JI, Reich T, Schmidt I, Schuckit MA (1994) A new, semi-
structured psychiatric interview for use in genetic linkage studies: a
report on the reliability of the SSAGA. J Stud Alcohol 55:149–158.
Buck KJ (1996) Molecular genetic analysis of the role of GABAergic
systems in the behavioral and cellular actions of alcohol. Behav Genet
Clayton D (1999) A generalization of the transmission/disequilibrium test
for uncertain haplotype transmission. Am J Hum Genet 65:1170–1177.
DICK ET AL.
Devlin B, Risch N (1995) A comparison of linkage disequilibrium mea- Download full-text
sures for fine-scale mapping. Genomics 29:311–322.
Dick DM, Foroud T (2003) Candidate genes for alcohol dependence: a
review of genetic evidence from human studies. Alcohol Clin Exp Res
Edenberg HJ, Dick DM, Xuei X, Foroud T, COGA collaborators (2002)
Genetic analysis of chromosome 4 for alcohol-related phenotypes. Am J
Med Genet 114:703.
Foroud T, Edenberg HJ, Goate A, Rice J, Flury L, Koller DL, Bierut LJ,
Conneally PM, Nurnberger JI, Bucholz KK, Li T-K, Hesselbrock V,
Crowe R, Schuckit M, Porjesz B, Begleiter H, Reich T (2000) Alcohol-
ism susceptibility loci: confirmation studies in a replicate sample and
further mapping. Alcohol Clin Exp Res 24:933–945.
Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B,
Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi
C, Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D
(2002) The structure of haplotype blocks in the human genome. Science
Ghosh S, Begleiter H, Porjesz B, Chorlian DB, Edenberg HJ, Foroud T,
Goate A, Reich T (2003) Linkage mapping of beta 2 EEG waves via
non-parametric regression. Am J Med Genet 118B:66–71.
Grobin AC, Matthews DB, Devaud LL, Morrow AL (1998) The role of
GABA-A receptors in the acute and chronic effects of ethanol. Psycho-
Hesselbrock M, Easton C, Bucholz KK, Schuckit M, Hesselbrock V (1999)
A validity study of the SSAGA—a comparison with the SCAN. Addic-
Martin ER, Bass MP, Kaplan NL (2001) Correcting for a potential bias in
the Pedigree Disequilibrium Test. Am J Hum Genet 68:1065–1067.
Martin ER, Monks SA, Warren LL, Kaplan NL (2000) A test for linkage
and association in general pedigrees: the Pedigree Disequilibrium Test.
Am J Hum Genet 67:146–154.
Meguro M, Mitsuya K, Sui H, Shigenami K, Kugoh H, Nakao M, Os-
himura M (1997) Evidence for uniparental, paternal expression of the
human GABAAreceptor subunit genes, using microcell-mediated chro-
mosome transfer. Hum Mol Genet 6:2127–2123.
Noble EP, Zhang X, Ritchie T, Lawford BR, Grosser SC, Young RM,
Sparkes RS (1998) D2 dopamine receptor and GABA-A receptor B3
subunit genes and alcoholism. Psychiatry Res 81:133–147.
Parsian A, Zhang Z-H (1999) Human chromosomes 11p15 and 4p12 and
alcohol dependence: possible association with the GABRB1 gene. Am J
Med Genet 88:533–538.
Porjesz B, Almasy L, Edenberg HJ, Wang K, Chorlian DB, Foroud T,
Goate A, Rice J, O’Connor SJ, Rohrbaugh J, Kuperman S, Bauer LO,
Crowe R, Schuckit M, Hesselbrock V, Conneally PM, Tischfield JA, Li
T-K, Reich T, Begleiter H (2002) Linkage disequilibrium between the
beta frequency of the human EEG and a GABAAreceptor gene locus.
Proc Natl Acad Sci USA 99:3729–3733.
Reich T (1996) A genomic survey of alcohol dependence and related
phenotypes: results from the Collaborative Study on the Genetics of
Alcoholism (COGA). Alcohol Clin Exp Res 20:133A–137A.
Reich T, Edenberg HJ, Goate A, Williams JT, Rice JP, Van Eerdewegh P,
Foroud T, Hesselbrock V, Schuckit MA, Bucholz K, Porjesz B, Li T-K,
Conneally PM, Nurnberger JI, Tischfield JA, Crowe RR, Cloninger CR,
Wu W, Shears S, Carr K, Crose C, Willig C, Begleiter H (1998)
Genome-wide search for genes affecting the risk for alcohol depen-
dence. Am J Med Genet 81:207–215.
Sobel E, Lange K (1996) Descent graphs in pedigree analysis: applications
to haplotyping, location scores, and marker sharing statistics. Am J
Hum Genet 58:1323–1337.
Song J, Koller DL, Foroud T, Rice J, Nurnberger JI Jr, Begleiter H,
Porjesz B, Smith TL, Schuckit M, Edenberg HJ (2003) Association of
GABA-A receptors and alcohol dependence and the effects of genetic
imprinting. Am J Med Genet 117B:39–45.
GABRG3 ASSOCIATED WITH ALCOHOLISM