How Can We Explain Very Low Odds Ratios in GWAS? I. Polygenic Models.

Genome-wide association studies of common diseases often identify a number of disease-related SNPs that reach highly significant p values but at the same time show very low disease odds ratios (ORs), most <1.5 and many <1.2. Despite their statistical significance, associations involving very low ORs explain little about the genetic contribution to the disease and nothing about disease inheritance. A commonly accepted explanation for very low ORs involves a model of polygenic inheritance, i.e., where the disease being studied is caused by a large number of interacting genes, each gene contributing only a small increment to disease risk. Here we demonstrate the perhaps counterintuitive result that, within a reasonable range of disease population prevalences (≤10%), a pure polygenic model is incompatible with very low ORs, unless very large numbers (hundreds or even thousands) of polygenic loci are involved.

Maximum Entropy Estimation of Probability Distribution of Variables in Higher Dimensions from Lower Dimensional Data.

A common statistical situation concerns inferring an unknown distribution Q(x) from a known distribution P(y), where X (dimension n), and Y (dimension m) have a known functional relationship. Most commonly, n ≤ m, and the task is relatively straightforward for well-defined functional relationships. For example, if Y1 and Y2 are independent random variables, each uniform on [0, 1], one can determine the distribution of X = Y1 + Y2; here m = 2 and n = 1. However, biological and physical situations can arise where n > m and the functional relation Y→X is non-unique. In general, in the absence of additional information, there is no unique solution to Q in those cases. Nevertheless, one may still want to draw some inferences about Q. To this end, we propose a novel maximum entropy (MaxEnt) approach that estimates Q(x) based only on the available data, namely, P(y). The method has the additional advantage that one does not need to explicitly calculate the Lagrange multipliers. In this paper we develop the approach, for both discrete and continuous probability distributions, and demonstrate its validity. We give an intuitive justification as well, and we illustrate with examples.

Causal models for investigating complex disease: I. A primer.

BACKGROUND/AIMS: To illustrate the utility of causal models for research in genetic epidemiology and statistical genetics. Causal models are increasingly applied in risk factor epidemiology, economics, and health policy, but seldom used in statistical genetics or genetic epidemiology. Unlike the statistical models usually used in genetic epidemiology, causal models are explicitly formulated in terms of cause and effect relationships occurring at the individual level. METHODS: We describe two causal models, the sufficient component cause model and the potential outcomes model, and show how key concepts in genetic epidemiology, including penetrance, phenocopies, genetic heterogeneity, etiologic heterogeneity, gene-gene interaction, and gene-environment interaction, can be framed in terms of these causal models. We also illustrate how potential outcomes models can provide insight into the potential for confounding and bias in the measurement of causal effects in genetic studies. RESULTS: Our analysis illustrates how causal models can elucidate the relationships among underlying causal mechanisms and measures obtained from statistical analysis of observed data. CONCLUSION: Causal models can enhance research aimed at identifying causal genes.

Causal models for investigating complex genetic disease: II. what causal models can tell us about penetrance for additive, heterogeneity, and multiplicative two-locus models.

BACKGROUND/AIMS: Statistical geneticists commonly use certain two-locus penetrance models because these models are familiar and mathematically tractable. We investigate whether and under what circumstances these two-locus penetrance models correspond to models of causation. METHODS: We describe a sufficient component cause model for a hypothetical disease with two genetic causes. We then use the potential outcomes framework to determine the expected two-locus penetrances from this causal model and contrast them with commonly used two-locus penetrance models (additive, heterogeneity, and multiplicative penetrance models, as formulated by Risch [Am J Hum Genet 1990;46:222-228]). RESULTS: Conventional additive and multiplicative models can correspond to any two-locus causal model only when certain very specific algebraic relationships hold. The heterogeneity model corresponds to a two-locus causal model only if the model stipulates that no disease cases are caused by the combined presence of the causal genotypes at both loci (i.e. only when there is no causal gene-gene interaction). Hence the heterogeneity model provides a valid test of the null hypothesis of no gene-gene interaction, whereas the additive and multiplicative models do not. CONCLUSION: We suggest that causal principles should provide the basis for statistical modeling in genetics.

Using parametric multipoint lods and mods for linkage analysis requires a shift in statistical thinking.

Multipoint (MP) linkage analysis represents a valuable tool for whole-genome studies but suffers from the disadvantage that its probability distribution is unknown and varies as a function of marker information and density, genetic model, number and structure of pedigrees, and the affection status distribution [Xing and Elston: Genet Epidemiol 2006;30:447-458; Hodge et al.: Genet Epidemiol 2008;32:800-815]. This implies that the MP significance criterion can differ for each marker and each dataset, and this fact makes planning and evaluation of MP linkage studies difficult. One way to circumvent this difficulty is to use simulations or permutation testing. Another approach is to use an alternative statistical paradigm to assess the statistical evidence for linkage, one that does not require computation of a p value. Here we show how to use the evidential statistical paradigm for planning, conducting, and interpreting MP linkage studies when the disease model is known (lod analysis) or unknown (mod analysis). As a key feature, the evidential paradigm decouples uncertainty (i.e. error probabilities) from statistical evidence. In the planning stage, the user calculates error probabilities, as functions of one's design choices (sample size, choice of alternative hypothesis, choice of likelihood ratio (LR) criterion k) in order to ensure a reliable study design. In the data analysis stage one no longer pays attention to those error probabilities. In this stage, one calculates the LR for two simple hypotheses (i.e. trait locus is unlinked vs. trait locus is located at a particular position) as a function of the parameter of interest (position). The LR directly measures the strength of evidence for linkage in a given data set and remains completely divorced from the error probabilities calculated in the planning stage. An important consequence of this procedure is that one can use the same criterion k for all analyses. This contrasts with the situation described above, in which the value one uses to conclude significance may differ for each marker and each dataset in order to accommodate a fixed test size, α. In this study we accomplish two goals that lead to a general algorithm for conducting evidential MP linkage studies. (1) We provide two theoretical results that translate into guidelines for investigators conducting evidential MP linkage: (a) Comparing mods to lods, error rates (including probabilities of weak evidence) are generally higher for mods when the null hypothesis is true, but lower for mods in the presence of true linkage. Royall [J Am Stat Assoc 2000;95:760-780] has shown that errors based on lods are bounded and generally small. Therefore when the true disease model is unknown and one chooses to use mods, one needs to control misleading evidence rates only under the null hypothesis; (b) for any given pair of contiguous marker loci, error rates under the null are greatest at the midpoint between the markers spaced furthest apart, which provides an obvious simple alternative hypothesis to specify for planning MP linkage studies. (2) We demonstrate through extensive simulation that this evidential approach can yield low error rates under the null and alternative hypotheses for both lods and mods, despite the fact that mod scores are not true LRs. Using these results we provide a coherent approach to implement a MP linkage study using the evidential paradigm.

An improved delta-centralization method for population stratification.

Dadd et al. [Hum Hered 2010;69:285-294] recently criticized our delta-centralization (DC) method of controlling for population stratification (PS) and concluded that DC does not work. To explore our method, the authors simulated data under the Balding-Nichols (BN) model, which is more general than the model we had used in our simulations. They determined that the DC method underestimated the PS parameter (δ) and inflated the type I error rates when applied to BN-simulated data, and from this they concluded that the DC method is invalid. However, we argue that this conclusion is premature. In this paper, we (1) show why δ is underestimated and type I error rates are inflated when BN-simulated data are used, and (2) present a simple adjustment to DC that works reasonably well for data from both kinds of simulations. We also show that the adjusted DC method has appropriate power under a range of scenarios.

Expected monotonicity--a desirable property for evidence measures?

We consider here the principle of 'evidential consistency' - that as one gathers more data, any well-behaved evidence measure should, in some sense, approach the true answer. Evidential consistency is essential for the genome-scan design (GWAS or linkage), where one selects the most promising locus(i) for follow-up, expecting that new data will increase evidence for the correct hypothesis. Earlier work [Vieland, Hum Hered 2006;61:144-156] showed that many popular statistics do not satisfy this principle; Vieland concluded that the problem stems from fundamental difficulties in how we measure evidence and argued for determining criteria to evaluate evidence measures. Here, we investigate in detail one proposed consistency criterion - expected monotonicity (ExpM) - for a simple statistical model (binomial) and four likelihood ratio (LR)-based evidence measures. We show that, with one limited exception, none of these measures displays ExpM; what they do display is sometimes counterintuitive. We conclude that ExpM is not a reasonable requirement for evidence measures; moreover, no requirement based on expected values seems feasible. We demonstrate certain desirable properties of the simple LR and demonstrate a connection between the simple and integrated LRs. We also consider an alternative version of consistency, which is satisfied by certain forms of the integrated LR and posterior probability of linkage.

The essence of linkage-based imprinting detection: comparing power, type 1 error, and the effects of confounders in two different analysis approaches.

Imprinting is critical to understanding disease expression. It can be detected using linkage information, but the effects of potential confounders (heterogeneity, sex-specific penetrance, and sex-biased ascertainment) have not been explored. We examine power and confounders in two imprinting detection approaches, and we explore imprinting-linkage interaction. One method (PP) models imprinting by maximising lod scores w.r.t. parent-specific penetrances. The second (DRF) approximates imprinting by maximising lods over differential male-female recombination fractions. We compared power, type 1 error, and confounder effects in these two methods, using computer-simulated data. We varied heterogeneity, penetrance, family and dataset size, and confounders that might mimic imprinting. Without heterogeneity, PP had more imprinting-detecting power than DRF. PP's power increased when parental affectedness status was ignored, but decreased with heterogeneity. With heterogeneity, type 1 error increased dramatically for both methods. However, DRF's power also increased under heterogeneity, more than was attributable to inflated type 1 error. Sex-specific penetrance could increase false positives for PP but not for DRF. False positives did not increase on ascertainment through an affected "mother". For PP, non-penetrant individuals increased information, arguing against using affected-only methods. The high type 1 error levels under some circumstances means these methods must be used cautiously.

The effect of inbreeding on the distribution of compound heterozygotes: a lesson from Lipase H mutations in autosomal recessive woolly hair/hypotrichosis.

Autozygosity mapping in consanguineous families has proven to be a powerful method for identifying recessive disease genes. Using this technique with whole genome SNP data generated from low density mapping arrays, we previously identified two genes that underlie autosomal recessive woolly hair (ARWH/hypotrichosis; OMIM278150), specifically P2RY5 and Lipase H (LIPH). In the current study, we sought to identify a novel disease locus for ARWH/hypotrichosis by analyzing two large consanguineous families from Pakistan who had initially been excluded for mutations at either of these disease loci by haplotype analysis with microsatellite markers. A genome-wide analysis of 10 members from each of the two families failed to identify significant regions of autozygosity or linkage. Upon genotyping an additional 10 family members in one of the families, parametric linkage analysis identified a region on chromosome 3q27 with evidence for linkage (Z = 2.5). Surprisingly, this region contains the LIPH gene. Microsatellite markers located within the LIPH gene were used for haplotype analysis and demonstrated that not one, but two haplotypes were segregating with the phenotype in each of these families. DNA sequencing identified two distinct LIPH mutations (280_369dup90 and 659_660delTA). Each affected individual (n = 38) was either homozygous for one mutation (n = 7 and 16 respectively), or compound heterozygous (n = 15). A review of the literature identified several reports of compound heterozygotes in consanguineous families. Prompted by this finding, we derived the probability that a patient affected with a recessive disease is carrying two mutations at the disease locus. We suggest that the validity of the IBD assumption may be challenged in large consanguineous families.

Multipoint lods provide reliable linkage evidence despite unknown limiting distribution: type I error probabilities decrease with sample size for multipoint lods and mods.

We investigate the behavior of type I error rates in model-based multipoint (MP) linkage analysis, as a function of sample size (N). We consider both MP lods (i.e., MP linkage analysis that uses the correct genetic model) and MP mods (maximizing MP lods over 18 dominant and recessive models). Following Xing and Elston (2006 Genet. Epidemiol, 30: 447-458), we first consider MP linkage analysis limited to a single position; then we enlarge the scope and maximize the lods and mods over a span of positions. In all situations we examined, type I error rates decrease with increasing sample size, apparently approaching zero. We show: (a) For MP lods analyzed only at a single position, well-known statistical theory predicts that type I error rates approach zero. (b) For MP lods and mods maximized over position, this result has a different explanation, related to the fact that one maximizes the scores over only a finite portion of the parameter range. The implications of these findings may be far-reaching: Although it is widely accepted that fixed nominal critical values for MP lods and mods are not known, this study shows that whatever the nominal error rates are, the actual error rates appear to decrease with increasing sample size. Moreover, the actual (observed) type I error rate may be quite small for any given study. We conclude that MP lod and mod scores provide reliable linkage evidence for complex diseases, despite the unknown limiting distributions of these MP scores.

A linkage search for joint panic disorder/bipolar genes.

There is comorbidity and a possible genetic connection between Bipolar disease (BP) and panic disorder (PD). Genes may exist that increase risk to both PD and BP. We explored this possibility using data from a linkage study of PD (120 multiplex families; 37 had > or =1 BP member). We calculated 2-point lodscores maximized over male and female recombination fractions by classifying individuals with PD and/or BP as affected (PD + BP). Additionally, to shed light on possible heterogeneity, we examine the pedigrees containing a bipolar member (BP+) separately from those that do not (BP-), using a Predivided-Sample Test (PST). Linkage evidence for common genes for PD + BP was obtained on chromosomes 2 (lodscore = 4.6) and chromosome 12 (lodscore = 3.6). These locations had already been implicated using a PD-only diagnosis, although at both locations this was larger when a joint PD + BP diagnosis was used. Examining the BP+ families and BP- families separately indicates that both BP+ and BP- pedigrees are contributing to the peaks on chromosomes 2 and 12. However, the PST indicates different evidence of linkage is obtained from BP+ and BP- pedigrees on chromosome 13. Our findings are consistent with risk loci for the combined PD + BP phenotype on chromosomes 2 and 12. We also obtained evidence of heterogeneity on chromosome 13. The regions on chromosomes 12 and 13 identified here have previously been implicated as regions of interest for multiple psychiatric disorders, including BP.

A unified approach for quantifying, testing and correcting population stratification in case-control association studies.

The HapMap project has given case-control association studies a unique opportunity to uncover the genetic basis of complex diseases. However, persistent issues in such studies remain the proper quantification of, testing for, and correction for population stratification (PS). In this paper, we present the first unified paradigm that addresses all three fundamental issues within one statistical framework. Our unified approach makes use of an omnibus quantity (delta), which can be estimated in a case-control study from suitable null loci. We show how this estimated value can be used to quantify PS, to statistically test for PS, and to correct for PS, all in the context of case-control studies. Moreover, we provide guidelines for interpreting values of delta in association studies (e.g., at alpha = 0.05, a delta of size 0.416 is small, a delta of size 0.653 is medium, and a delta of size 1.115 is large). A novel feature of our testing procedure is its ability to test for either strictly any PS or only 'practically important' PS. We also performed simulations to compare our correction procedure with Genomic Control (GC). Our results show that, unlike GC, it maintains good Type I error rates and power across all levels of PS.

A third-pass genome scan in panic disorder: evidence for multiple susceptibility loci.

BACKGROUND: Panic disorder (PD) is a common illness with a definite but "complex" genetic contribution and estimated heritability of 30-46%. METHODS: We report a genome scan in 120 multiplex PD pedigrees consisting of 1591 individuals of whom 992 were genotyped with 371 markers at an average spacing of 9cM. Parametric two-point, multipoint, and nonparametric analyses were performed using three PD models (Broad, Intermediate, Narrow) and allowing for homogeneity or heterogeneity. The two-point analyses were also performed allowing for independent male and female recombination fractions (theta). Genome-wide significance was empirically evaluated using simulations of this dataset. RESULTS: Evidence for linkage reached genome-wide significance in one region on chromosome 15q (near GABA-A receptor subunit genes) and was suggestive at loci on 2p, 2q and 9p using an averaged theta. Analyses allowing for sex-specific theta's were consistent except that support at one locus on 2q increased to genome-wide significance and an additional region of suggestive linkage on 12q was identified. However, differences in male and female recombination fractions predicted by the sex-specific approach were not consistent with current physical maps. CONCLUSIONS: These data provide evidence for chromosomal regions on 15q and 2q that may be important in genetic susceptibility to panic disorder. Although we are encouraged by the findings of analyses using sex-specific recombination fractions, we also note that further understanding of this analytic strategy will be important.

Possible interaction between HLA-DRbeta1 and thyroglobulin variants in Graves' disease.

Graves' disease (GD) is influenced by two major susceptibility loci, HLA-DR3 and thyroglobulin (Tg). Recently we have shown that specific HLA-DR and Tg gene sequences predispose to Graves' disease. Individuals carrying at least one arginine at position 74 of the DRbeta1 chain (denoted the R- genotype) have a significantly increased risk of GD, as do individuals homozygous for the single nucleotide protein (SNP) in exon 33 of the Tg gene (denoted the CC genotype). Therefore, for the current study we hypothesized that these two genes may interact to influence the etiology of GD. To test this hypothesis, we analyzed the genotypes of 185 Caucasian patients with GD and 143 Caucasian controls for both genes. We tested for an interaction effect, that is, is one gene's effect on GD greater when the other gene is also present than when the other gene is absent? A logistic regression analysis yielded an estimate of 4.31 for the interaction term (p = 0.053). Our results may suggest an interaction between the R- and CC variants in conferring susceptibility to GD. These results, if confirmed, may imply that these two variants interact biologically to increase the odds of GD.

Using Linkage Analysis to Detect Gene-Gene Interactions. 2. Improved Reliability and Extension to More-Complex Models.

Detecting gene-gene interaction in complex diseases has become an important priority for common disease genetics, but most current approaches to detecting interaction start with disease-marker associations. These approaches are based on population allele frequency correlations, not genetic inheritance, and therefore cannot exploit the rich information about inheritance contained within families. They are also hampered by issues of rigorous phenotype definition, multiple test correction, and allelic and locus heterogeneity. We recently developed, tested, and published a powerful gene-gene interaction detection strategy based on conditioning family data on a known disease-causing allele or a disease-associated marker allele4. We successfully applied the method to disease data and used computer simulation to exhaustively test the method for some epistatic models. We knew that the statistic we developed to indicate interaction was less reliable when applied to more-complex interaction models. Here, we improve the statistic and expand the testing procedure. We computer-simulated multipoint linkage data for a disease caused by two interacting loci. We examined epistatic as well as additive models and compared them with heterogeneity models. In all our models, the at-risk genotypes are "major" in the sense that among affected individuals, a substantial proportion has a disease-related genotype. One of the loci (A) has a known disease-related allele (as would have been determined from a previous analysis). We removed (pruned) family members who did not carry this allele; the resultant dataset is referred to as "stratified." This elimination step has the effect of raising the "penetrance" and detectability at the second locus (B). We used the lod scores for the stratified and unstratified data sets to calculate a statistic that either indicated the presence of interaction or indicated that no interaction was detectable. We show that the new method is robust and reliable for a wide range of parameters. Our statistic performs well both with the epistatic models (false negative rates, i.e., failing to detect interaction, ranging from 0 to 2.5%) and with the heterogeneity models (false positive rates, i.e., falsely detecting interaction, ≤1%). It works well with the additive model except when allele frequencies at the two loci differ widely. We explore those features of the additive model that make detecting interaction more difficult. All testing of this method suggests that it provides a reliable approach to detecting gene-gene interaction.

Interstitial cystitis and panic disorder: a potential genetic syndrome.

BACKGROUND: Evidence from a genetic linkage study had suggested a possible syndrome in some families with panic disorder (PD). This syndrome includes bladder problems (possibly urinary interstitial cystitis [IC]), thyroid disorders, chronic headaches/migraine, and/or mitral valve prolapse. In 19 multiplex families with PD, one marker (D13S779) on chromosome 13 gave a logarithm of odds score of more than 4 when individuals with any of the syndrome conditions were analyzed as affected. Families with the bladder problems yielded the highest logarithm of odds scores. These findings were replicated in an extended sample of 60 families. Whereas PD had been well characterized by direct interview, the urologic problems had been found only via medical history checklists and records. A case review by a board-certified urologist suggested they could be IC. OBJECTIVE: To determine whether patients diagnosed as having IC by urodynamics and/or cystoscopy and their first-degree relatives (FDRs) have increased rates of the syndrome conditions, thus validating that the bladder problems observed in the linkage study could be IC and providing further support for the panic syndrome. DESIGN: Case-control and family history study. SETTING: Two metropolitan urology clinics. PARTICIPANTS: One hundred forty-six probands (67 with IC and 79 with other urologic disorders) and 815 FDRs. MAIN OUTCOME MEASURES: Lifetime rates of syndrome conditions in probands and FDRs who were blind to urologic or psychiatric diagnoses in the proband. RESULTS: Compared with patients without IC, patients with IC had a significantly higher lifetime prevalence of PD (controlling for age and sex) (odds ratio, 4.05; 95% confidence interval, 1.22-13.40; P =.02) and a higher lifetime prevalence of any of the syndrome disorders (controlling for age and sex) (odds ratio, 2.22; 95% confidence interval, 0.89-5.54; P =.09). First-degree relatives of probands with (vs without) IC were significantly more likely to have PD, thyroid disorder, urologic problems, and any of the syndrome disorders (controlling for age and sex of the relative and sex of the proband) (adjusted odds ratio, 1.95; 95% confidence interval, 1.13-3.38; P =.02). These results in relatives were not influenced by PD in probands, and did not change substantially when controlling for the proband-relative relationship, modeling age as a categorical (vs continuous) variable, or excluding FDRs with PD. There were no interactions between proband IC status and sex of the relative. CONCLUSIONS: The increased frequency of seemingly disparate disorders in patients with IC and their FDRs is consistent with the genetic linkage findings in families with PD. These findings suggest that the bladder problems observed in the linkage study may be IC. The hypothesis that there is a familial, possibly pleiotropic, syndrome that may include IC, PD, thyroid disorders, and other disorders of possible autonomic or neuromuscular control deserves further investigation.

Evidence for a susceptibility locus for panic disorder near the catechol-O-methyltransferase gene on chromosome 22.

BACKGROUND: A well-characterized single nucleotide polymorphism (472G/A-Val/Met-SNP8) in the coding sequence of the catechol-O-methyltransferase (COMT) gene leads to a three- to fourfold difference in enzymatic activity and clinical and animal studies suggest a role in anxiety states like panic disorder. METHODS: Subjects from 70 panic disorder pedigrees, and 83 "triads", were genotyped at seven single nucleotide polymorphisms (SNPs), polymorphic microsatellites in the first intron of COMT and approximately 339kb upstream of COMT (D22S944) and analyzed for genetic association and linkage. RESULTS: Linkage analysis showed elevated LOD scores for 472G/A (SNP 8), silent exon 3 substitution (186C/T-SNP 5), and the marker D22S944 (2.88, 2.62, and 2.93, respectively), using a variety of diagnostic and genetic models. Association tests were not significant for the SNPs, but were highly significant for D22S944 (p =.0001-.0003). One three-marker haplotype formed from the above three polymorphisms was significantly associated with panic disorder (p =.0001), as was the "global" p value for this combination (p =.005). In addition, numerous haplotypes with combinations of D22S944 and COMT SNPs were found to be significantly associated with panic disorder. CONCLUSIONS: Our findings provide strong evidence for a susceptibility locus for panic disorder either within the COMT gene or in a nearby region of chromosome 22.

Evidence for genetic linkage between a polymorphism in the adenosine 2A receptor and panic disorder.

Data from clinical and behavioral pharmacological studies have implicated adenosine in anxiety behaviors, while genetic studies have suggested that adenosine receptors may be associated with panic disorder. We have undertaken an analysis of several DNA sequence variations in the adenosine 2A receptor (ADORA2A) in a large sample of panic disorder pedigrees. Individuals from 70 panic disorder pedigrees, and 83 child-parent 'trios', were genotyped at five single-nucleotide polymorphisms (SNPs) in and near the ADORA2A gene and were analyzed for genetic linkage and association. Linkage analysis revealed elevated LOD scores for a silent substitution (1083C/T, SNP-4) in the second coding exon. This SNP has been previously reported to be associated with panic disorder. We observed a maximal heterogeneity LOD score of 2.98 (theta=0) under a recessive genetic model and narrow diagnostic model. Other SNPs showed no evidence for linkage. Association tests were not significant for any of the five ADORA2A SNPs. When SNP haplotypes were assessed in the triads with TRANSMIT, one 3-marker haplotype (SNPs 1, 4, 5) was nominally significantly associated with panic disorder (p=0.029). Pairwise estimations of linkage disequilibrium between the SNPs showed strong patterns of linkage disequilibrium across the ADORA2A locus. Analyses carried out by broadening the panic disorder phenotype to include agoraphobia continued to support linkage to ADORA2A. Our findings provide evidence for a susceptibility locus for panic disorder, and possibly including agoraphobia, either within the ADORA2A gene or in a nearby region of chromosome 22, and serves as the first successful candidate gene replication study in panic disorder.