Covariates in linkage analysis using sibling and cousin pairs

Using simulated data from GAW 12, problem 2, we further develop a novel technique to detect and use significant covariates in linkage analysis. The method, first introduced by Rice et al. [Genet Epidemiol 17(Suppl. 1):S691-5, 1999], uses logistic regression to model perturbation in sharing as a function of covariate levels. The original method allows use of all sib pairs (concordant affected, concordant unaffected, and discordant). Here we extend this method to include cousin pairs in analysis.

A genome screen of maximum number of drinks as an alcoholism phenotype

The Collaborative Study on the Genetics of Alcoholism (COGA) is a multicenter research program to detect and map susceptibility genes for alcohol dependence and related phenotypes. The measure M of "maximum number of drinks consumed in a 24-hour period" is closely related to alcoholism diagnosis in this dataset and provides a quantitative measure to grade nonalcoholic individuals. Twin studies have shown log(M) to have a heritability of approximately 50%. Genome screens for this trait were performed in two distinct genotyped samples (wave 1 and wave 2), and in the combined sample. MAPMAKER/SIBS was used to carry out Haseman-Elston based regression analyses. On chromosome 4, an unweighted all-pairs multipoint LOD of 2.2 was obtained between D4S2407 and D4S1628 in wave 1; in wave 2, the region flanked by D4S2404 and D4S2407 gave a LOD of 1.5. In the combined sample, the maximal LOD was 3.5 very close to D4S2407. This evidence for linkage is in the region of the alcohol dehydrogenase gene cluster on chromosome 4. These findings on chromosome 4 are consistent with a prior report from COGA in which strictly defined nonalcoholic subjects in wave 1 were analyzed. The present analysis on log(M) allows more individuals to be included and thus is potentially more powerful.

The utility of deviant sib pairs in the detection of linkage

A tripartite sampling design was used to help deduce the genetic structure of a complex biological system. Univariate and multivariate population parameters were estimated from an age/sex stratified sample of unrelated individuals. Estimates of familial resemblance between and within continuous variables were obtained from a sample of randomly ascertained nuclear families. Finally, a sample of highly deviant concordant and discordant independent sib pairs facilitated the discovery of major genes through multipoint linkage analysis. No false positive signals were inferred. The pleiotropic effects of major genes, however, became obscured when linkage analysis was performed on adjusted quantitative variables.

Covariates in linkage analysis

We apply a novel technique to detect significant covariates in linkage analysis using a logistic regression approach. An overall test of linkage is first performed to determine whether there is significant perturbation from the expected 50% sharing under the hypothesis of no linkage; if the overall test is significant, the importance of the individual covariate is assessed. In addition, association analyses were performed. These methods were applied to simulated data from multiple populations, and detected correct marker linkages and associations. No population heterogeneity was detected. These methods have the advantages of using all sib pairs and of providing a formal test for heterogeneity across populations.

Genome screen for platelet monoamine oxidase (MAO) activity

To identify loci involved in the control of platelet monoamine oxidase B (MAO-B) activity, a genomewide linkage screen was performed using 291 markers in 148 nuclear families containing a total of 1,008 nonindependent sib-pairs. Participants were genotyped and their platelet MAO-B activity levels were measured as part of the Collaborative Study on the Genetics of Alcoholism (COGA). Sib-pair analysis using Haseman-Elston regression was carried out with two programs. Two-point analysis on all pairs with SIBPAL indicated three markers with p-values below 0.01: D6S1018 (p = 0.0004), D2S1328 (p = 0.008), and D2S408 (p = 0.003). MAPMAKER/SIBS multipoint analyses using independent pairs(N = 409) gave maximal lod scores of 2. 0 on chromosome 6 and 1.1 and 1.4 for the two regions on chromosome 2. These results are consistent with linkage, but do not provide definitive evidence. We are currently creating a denser map in these regions and have begun genotyping a second sample in COGA.

Comparison of direct interview and family history diagnoses of alcohol dependence

Using data from The Collaborative Study on the Genetics of Alcoholism, we compare direct interview diagnoses of alcohol dependence to those obtained by history from family members. Using a requirement of three or more positive implications by history, the specificity, sensitivity, and positive predictive values are 98%, 39%, and 45%, respectively. A logistic analysis found the gender of the relative and alcoholism in the informant to be significant, but not the gender of the informant. The partial odds ratio of a diagnosis at interview associated with a positive family history diagnosis was 13.6. The relationship between the informant and relative was significant, with negative reports from an offspring or mate more influential than a negative report from a parent or second-degree relative. We derived a recursive equation to combine a variable number of family history reports, wherein the probabilities associated with a single report are computed from the logistic analysis. This permits the use of family history information both as a proxy for an uninterviewed relative, as well as a second source of information to be used in the analysis of genetic family data.

Affective illness in family members and matched controls

As part of the US National Institute of Mental Health Collaborative Program on the Psychobiology of Depression study, a subset of 460 randomly chosen relatives of affectively ill probands were compared to a control group matched by the acquaintanceship method. The rate of major affective disorder in relatives was found to be 36%; the rate among controls was 28%. Relatives were also found to have significantly higher rates of bipolar II disorder, any Research Diagnostic Criteria (RDC) affective disorder and any RDC mental disorder. All of these rates were found to be significantly higher when female relatives were compared with their acquaintances, but only the rate of any RDC mental disorder was higher when this comparison was made in men. The acquaintanceship method enabled the selection of a control group that closely resembled the relatives, probably to the extent of "overmatching". When the match was evaluated to determine whether relatives tended to select comparably ill (or well) acquaintances, this was found to be the case only for alcoholic and never mentally ill relatives.

Stability of psychiatric diagnoses. An application to the affective disorders

In the National Institute of Mental Health Collaborative Program on the Psychobiology of Depression study, data were collected on 2226 first-degree relatives of 612 probands. A second, "blind" reassessment of all relatives was attempted 6 years after the initial evaluation. We report on a final sample of 1629 relatives assessed twice using the Schedule for Affective Disorders and Schizophrenia-Lifetime version. We summarize methods for using stability of diagnosis to model the relationship between clinical covariates and the probability of being a true case. Moreover, we define an index of caseness that can be used to narrow the criteria for who is a case. Of those positive for major depressive disorder at initial evaluation, 74% were positive (on a lifetime basis) at follow-up (ie, were stable). There is a gradient: 48% of those who had three symptoms and no treatment were stable, compared with 96% of those with eight symptoms and treatment. For major depressive disorder, we found the caseness index for those with lifetime mania more severe than that of nonbipolar patients, with those who had hypomania being intermediate. A hierarchical analysis indicated that bipolar I tends to be diagnosed as schizoaffective-manic across occasions, and vice versa. This is consistent with the prior familial analyses that suggest these two diagnoses be combined into a single bipolar phenotype. The analysis for major depressive disorder indicates that caseness appears to represent quantitative, rather than qualitative, differences, with no natural cutoff to identify distinct subgroups. Finally, we discuss implications including utility in genetic analyses, estimation of incidence or prevalence allowing for diagnostic error, and examination of cohort effects.

Schizo-affective disorders: bipolar-unipolar subtyping. Natural history variables: a discriminant analysis approach

In view of tackling the problem of heterogeneity among the schizo-affectives, methods of univariate and multivariate statistical analysis (canonical discriminant analysis) were applied to the sociodemographic and natural history variables of four groups of affective disorder patients from the NIMH Collaborative Study on the Psychobiology of Depression Clinical section: the schizo-bipolar (SBP, n = 45), the schizo-unipolar (SUP, n = 30), the bipolar I (BP, n = 159) and the primary unipolar depressed (UP, n = 387) defined by Research Diagnostic Criteria. Two dimensions were identified among the four groups of 'affective' patients: the 'bipolar' and the 'schizophrenic' dimensions. They provided highly significant discrimination among the means of the four groups but were not very accurate in predicting group membership. The 'bipolar' dimension separates the UP from the BP and SBP, the SUP taking some intermediate value. The 'schizophrenic' dimension separates the BP and UP from the SUP, the SBP being intermediate. The two groups with the most similarities were the SBP and BP. The group with the most heterogeneity was the SUP, sharing similarities with the UP and SBP mostly. These conclusions are supported by results of familial aggregation on the same group of patients.

The antisocial and the nonantisocial alcoholic: clinical distinctions in men with major unipolar depression

Depressed, alcoholic men were divided into those with and those without antisocial personality (ASP). The ASP alcoholics had a more severe, cross-sectional clinical picture; they had a higher rate of drug use disorder and more alcohol-related symptomatology. The medical complications of alcoholism were related to the duration of heavy drinking whereas the nonmedical complications (the psychosocial and pathologic-intake problems) were related to ASP, duration of heavy drinking, and their interaction.

The estimation of diagnostic sensitivity using stability data: an application to major depressive disorder

A difficulty in the interpretation of the reliability/stability of a lifetime diagnosis of mental disorders is the lack of a theoretical perspective. A model expressed in terms of the three unknowns--sensitivity, specificity and true base rate--is problematic due to the lack of a "gold standard", so that only two of these unknowns can be estimated. We extend this model to allow for clinical covariates that increase the likelihood that a positive case at time 1 will be positive at time 2. Under the assumption that all observed cases are true cases at the highest covariate values, we obtain a direct estimate of the sensitivity, so that all unknowns can be estimated. Moreover, we then calculate the likelihood that an observed case with given covariate levels is in fact a true case. The implications of diagnostic error for the estimation of incidence and population rates and the fitting of genetic models are given. These methods are applied to stability data collected as part of the NIMH Psychobiology of Depression Program. A total of 982 relatives have been assessed with interviews separated by a 6 year interval. A logistic function was used to model the stability in 264 relatives with an initial lifetime diagnosis of Major Depressive Disorder. The significant covariate predictors in a step-wise analysis were the number of depressive symptoms in the worst episode, the number of episodes and receiving medication for the worst episode. The sensitivity was estimated to be 0.918 and the specificity was estimated to be 0.940.(ABSTRACT TRUNCATED AT 250 WORDS)