Commingling and segregation analyses: comparison of results from a simulation study of a quantitative trait

Evidence of Commingling in Human Eating Behavior

This investigation tested whether distributions of certain aspects of eating behavior were consistent with the notion of a "mixture model;" that is, two or more distinct commingled component distributions, consistent with the possibility of major gene action. Undergraduates (n=901) completed self-report trait measures of hunger, disinhibition, and dietary restraint. Variables were residualized for gender and age and transformed to remove skewness. Residualized transformed distributions were tested for departure from unimodality with Hartigan's dip statistic. The distributions of all three aspects of eating behavior were significantly non-unimodal. Next, component multivariate normal distributions were estimated via maximum likelihood. Likelihood ratio tests were employed to compare nested models. A mixture of four distributions with unequal variance-covariance matrices fit significantly better than any more parsimonious model. In sum, these data strongly suggest that the distributions of several measures of eating behavior are composed of four component distributions. This finding is consistent with the possibility of major gene effects for eating behavior.

Finding suitable phenotypes for genetic studies of schizophrenia: heritability and segregation analysis

BACKGROUND

Schizophrenia is a highly heritable and complex disorder. Multiple genes are likely to be involved, complicating genetic research into the etiology of this disorder. Intermediate phenotypes or endophenotypes may facilitate genetic research if they display a simpler mode of transmission than schizophrenia itself, i.e., if they reflect more closely the underlying genetic effects.

METHODS

Twenty-five multigenerational families with multiple members affected with schizophrenia (180 subjects) were administered an extensive neuropsychological, psychophysiological, and personality test battery. Familial correlations were calculated to select heritable traits. Subsequent heritability analysis followed by commingling and segregation analysis were performed to unravel the pattern of transmission and to estimate heritability.

RESULTS

Five traits, including sensorimotor gating, openness, verbal fluency, early visual perception, and spatial working memory, showed moderate familial correlations. Heritability estimates for these traits ranged from 37% to 54%. A major gene model resembling dominant transmission was found for both sensorimotor gating and openness. Verbal fluency, early visual perception, and spatial working memory may be accounted for by polygenic, multifactorial, or environmental effects.

CONCLUSIONS

Only 2 of 13 candidate endophenotypes showed a simple mode of transmission useful for successful application in molecular genetic research: sensorimotor gating and openness. To our knowledge, this is the first study to investigate the pattern of transmission for these traits.

Evidence for a Major Gene Influence on Tumor Necrosis Factor-α Expression in Tuberculosis: Path and Segregation Analysis

OBJECTIVE

Tuberculosis (TB) is a growing global public health problem. Several studies suggest a role for host genetics in disease susceptibility, but studies to date have been inconsistent and a comprehensive genetic model has not emerged. A limitation of previous genetic studies is that they only analyzed the binary trait TB, which does not reflect disease heterogeneity. Furthermore, these studies have not accounted for the influence of shared environment within households on TB risk, which may spuriously inflate estimates of heritability.

METHODS

We conducted a household contact study in a TB-endemic community in Uganda. Antigen-induced tumor necrosis factor-alpha (TNFalpha) expression, a key component of the underlying immune response to TB, was used as an endophenotype for TB.

RESULTS

Path analysis, conducted to assess the effect of shared environment, suggested that TNFalpha is heritable (narrow sense heritability = 34-66%); the effect of shared environment is minimal (1-14%), but gene-environment interaction may be involved. Segregation analysis of TNFalpha suggested a major gene model that explained one-third of the phenotypic variance, and provided putative evidence of natural selection acting on this phenotype.

CONCLUSION

Our data further support TNFalpha as an endophenotype for TB, as it may increase power to detect disease-predisposing loci.

Commingling Analysis of Intraocular Pressure and Glaucoma in an Older Australian Population

PURPOSE

To test the hypothesis of a major genetic determinant of intraocular pressure (IOP) under the mixed genetic model in a defined population, and to elucidate the relationship between IOP and glaucoma.

METHODS

IOP was measured in 3654 persons attending the Blue Mountains Eye Study. A commingling analysis on IOP was performed using a new program, SKUDRIVER. The goodness of fit of 1-, 2- and 3-distribution models was measured. This was repeated after the glaucoma cases had been removed from the dataset, and further repeated on the glaucoma cases in the dataset.

RESULTS

The best model was 3-distribution with no evidence for Hardy-Weinberg disequilibrium. The proportion of variance explained by this major effect was 0.18. When glaucoma cases were removed, the best model had 2 distributions. There was no evidence of admixture in glaucoma patients.

CONCLUSIONS

The findings of this study are consistent with the presence of a major gene accounting for 18% of the variance of IOP, which could therefore influence the risk of glaucoma, in this population. These findings suggest methods of optimizing strategies for family and association studies to identify quantitative trait loci for IOP. No evidence for distinct IOP-dependent and IOP-independent subgroups of glaucoma was found.

Consistency of genetic inheritance mode and heritability patterns of triglyceride vs. high density lipoprotein cholesterol ratio in two Taiwanese family samples

BACKGROUND

Triglyceride/HDL cholesterol ratio (TG/HDL-C) is considered as a risk factor for cardiovascular events. Genetic components were important in controlling the variation in western countries. But the mode of inheritance and family aggregation patterns were still unknown among Asian-Pacific countries. This study, based on families recruited from community and hospital, is aimed to investigate the mode of inheritance, heritability and shared environmental factors in controlling TG/HDL-C.

RESULTS

Two populations, one from community-based families (n = 988, 894 parent-offspring and 453 sibling pairs) and the other from hospital-based families (n = 1313, 76 parent-offspring and 52 sibling pairs) were sampled. The population in hospital-based families had higher mean age values than community-based families (54.7 vs. 34.0). Logarithmic transformed TG/ HDL-C values, after adjusted by age, gender and body mass index, were for genetic analyses. Significant parent-offspring and sibling correlations were also found in both samples. The parent-offspring correlation coefficient was higher in the hospital-based families than in the community-based families. Genetic heritability was higher in community-based families (0.338 +/- 0.114, p = 0.002), but the common shared environmental factor was higher in hospital-based families (0.203 +/- 0.042, p < 0.001). Commingling analyses showed that more than one-component distribution models were the best-fit models to explain the variance in both populations. Complex segregation analysis by regressive models revealed that in both samples the best-fit model of TG/HDL-C was the model of environmental effects plus familial correlation, in which significant parent-offspring and sibling correlations were demonstrated. Models of major gene effects were rejected in both samples.

CONCLUSION

Variations of TG/HDL-C in the normal ranges were likely to be influenced by multiple factors, including environmental and genetic components. Higher genetic factors were proved in younger community-based families than in older hospital-based families.

Major gene effects in systolic and diastolic blood pressure in families receiving a health examination in Taiwan

BACKGROUND

Genetic factors play an important role in determining blood pressure values. Strong familial aggregation and estimated heritability values around 0.4-0.6 were reported in a white population.

OBJECTIVES

To investigate the genetic components and mode of inheritance of blood pressure in Taiwan.

DESIGN

A cross-sectional family study based on an adult population undergoing a health examination.

SETTING AND PARTICIPANTS

A total of 1313 adults and their spouses or first-degree relatives, or both, underwent a health examination in a tertiary university hospital from August 1998 to September 1999.

MAIN OUTCOME MEASURES

Genetic analyses, including comingling analysis, familial correlation and complex segregation analysis, were used to detect the genetic components and the mode of inheritance of blood pressures.

RESULTS

In both systolic and diastolic blood pressure, the comingling analyses demonstrated that a more than one-component distribution provided the best fit for the data. The familial correlation coefficients showed significant parent-offspring and sibling correlation. Complex segregation analyses showed major gene effects in controlling systolic and diastolic blood pressure.

CONCLUSIONS

There were significant genetic components in blood pressure among Taiwanese. Further investigation of genomic loci for the control of blood pressure is indicated.

Genetic contributions to LDL-C, Apo-B and LDL-C/Apo-B ratio in a sample of Israeli offspring with a parental history of myocardial infarction

One hundred and forty sibships consisting of 280 brothers and 256 sisters with a family history of myocardial infarction were investigated for the possible involvement of a major gene in the determination of LDL-C, Apo-B and LDL-C/ Apo-B ratio (as a surrogate for LDL subclasses). The mean ages were 29.5 years (range 15-48) and 29.2 years (range 15-47), for brothers and sisters, respectively, and values were initially adjusted for age effects through multiple regression analysis. Results from commingling analysis indicated that for LDL-C a single normal distribution fitted the data as well as a mixture of two distributions. For Apo-B, a mixture of two normal distributions fitted the data significantly better than a single distribution (chi 2 = 7.8, df = 2, p = 0.02). For LDL-C/ Apo-B ratio a mixture of three normal distributions fitted the data significantly better than two distributions (chi 2 = 9.2, df = 2, p = 0.01). A regression analysis applied to the logarithm of the sex- and age-adjusted mean and variance within sibship, showed no indication of a major gene involvement for LDL-C. For Apo-B and LDL-C/Apo-B ratio, there existed, however, significant linear relationships between the logarithmically transformed means and within sibship variances which support the involvement of major genes. In addition, the Bartlett test applied to the data of within-sibship variances also rejected the null hypothesis of multifactorial transmission for Apo-B and LDL-C/Apo-B ratio (p < 0.0001). Lastly, the results from sib-pair linkage analyses provided significantly positive evidence for linkage between ApoB levels and the Apo-B XbaI restriction site.

Mixture distributions in human genetics research

The use of mixture distributions in genetics research dates back to at least the late 1800s when Karl Pearson applied them in an analysis of crab morphometry. Pearson's use of normal mixture distributions to model the mixing of different species of crab (or 'families' of crab as he referred to them) within a defined geographic area motivated further use of mixture distributions in genetics research settings, and ultimately led to their development and recognition as intuitive modelling devices for the effects of underlying genes on quantitative phenotypic (i.e. trait) expression. In addition, mixture distributions are now used routinely to model or accommodate the genetic heterogeneity thought to underlie many human diseases. Specific applications of mixture distribution models in contemporary human genetics research are, in fact, too numerous to count. Despite this long, consistent and arguably illustrious history of use, little mention of mixture distributions in genetics research is made in many recent reviews on mixture models. This review attempts to rectify this by providing insight into the role that mixture distributions play in contemporary human genetics research. Tables providing examples from the literature that describe applications of mixture models in human genetics research are offered as a way of acquainting the interested reader with relevant studies. In addition, some of the more problematic aspects of the use of mixture models in genetics research are outlined and addressed.

Commingling analysis of the distribution of a phenotype conditioned on two marker genotypes: application to plasma angiotensin-converting enzyme levels

Commingling analysis is a statistical method for distinguishing between one (usually normal) distribution and a mixture of two or more distributions. It is used in genetic studies, since one mechanism giving rise to a mixture of distributions is the effect of a major gene. Plasma levels of angiotensin-converting enzyme (ACE) have been shown to be related to an insertion/deletion (I/D) polymorphism in the ACE gene. Recently, Cambien et al. [(1994) Circulation 90:669-676] used commingling analysis in a study of ACE, extending the method to condition on information about the I/D polymorphism in the gene. A further common polymorphism has been discovered recently by our group [Foy et al. (1995) Blood Coagulation Fibrinolysis 6:590] in the promoter region of the gene. In this paper we extend the method of commingling analysis to condition on two marker loci. The method is illustrated by application to plasma ACE levels in subjects for whom genetic information at both the I/D and promoter loci has been recorded. The results confirm strong evidence for a mixture of three normal distributions in preference to one or two, and also show that neither polymorphism is identical to a putative functional polymorphism. To assess the performance of the method, a simulation study was carried out. Parameters may be estimated more efficiently conditioning on two marker loci than on one or none, but this is not always the case if the underlying distributions are well separated. Conditioning on marker loci can increase the power to distinguish between alternative hypotheses of interest.

Commingling and segregation analysis of reading performance in families of normal reading probands

This paper reports the results of commingling and genetic segregation analyses performed on a quantitative reading phenotype in 125 families ascertained through normal, nondisabled readers. Commingling analysis using SKUMIX suggested that the reading phenotype best fit a skewed, single distribution model. Complex segregation using POINTER was then performed on the power adjusted data. While there were some analytical ambiguities and complexities, the segregation analysis indicated that there was familial transmission of the phenotype and that a significant percentage of the variance in this phenotype could be attributed to a major gene with dominance. Because the estimated frequency of the putative dominant allele is .35, 57% of the population would carry at least one copy of this allele. This common allele, with low penetrance, accounted for 54% of the phenotypic variance in reading scores. These findings are considered in the context of our earlier report of major gene influence ona qualitative dyslexic phenotype in a sample of 133 dyslexic proband families that were originally matched to the present sample of control families (Pennington et al., 1991). The applicability of a classic single gene, multifactorial-polygenic, and oligogenic or QTL models for reading ability/disability is discussed.

Genetics of LDL subclass phenotypes in women twins. Concordance, heritability, and commingling analysis

Low density lipoprotein (LDL) subclass phenotype B, characterized by a predominance of small LDL as determined by gradient gel electrophoresis, has been associated with increased risk of coronary heart disease and an atherogenic lipoprotein profile. Previous studies employing complex segregation analysis have demonstrated a major, single gene effect on the inheritance of this phenotype in families. Recently, linkage between this phenotype and variation at the LDL receptor locus on chromosome 19 has been reported. However, variation in LDL subclass phenotypes has also been associated with age, gender, diabetes status, beta-blocker medication, and diet. The present study further evaluates the relative importance of genetic and nongenetic influences on LDL subclass phenotypes and on LDL peak particle diameter (as a reflection of the size of the major LDL subclass) in monozygotic and dizygotic women twin pairs. The analysis is based on 203 monozygotic and 145 dizygotic pairs of adult female twins who participated in the second examination of the Kaiser Permanente Women Twins Study. The average age was 51 years at this exam and 90% were white. Concordance analysis revealed that monozygotic cotwins shared LDL subclass phenotypes more frequently than dizygotic cotwins, and this was confirmed using logistic regression analysis after controlling for potential confounding factors. Heritability analyses suggested that approximately one third to one half of the variation in LDL peak particle diameter, a continuous variable reflecting LDL size, could be attributed to genetic influences. Commingling analysis of the frequency distribution of LDL peak particle diameter identified three distinct subgroups of subjects, one of which corresponded to those subjects with LDL subclass phenotype B.(ABSTRACT TRUNCATED AT 250 WORDS)

Detecting heterogeneity with the affected-pedigree-member (APM) method

We have developed a novel test of heterogeneity based on the APM method. We present tests of this method on both empiric and simulated data.

Bimodality of plasma apolipoprotein B levels in familial combined hyperlipidemia

To investigate possible genetic influences on plasma apolipoprotein (apo) B levels in familial combined hyperlipidemia (FCHL), commingling analysis was performed on data from seven large kindreds, including 183 individuals. The overall frequency distribution of apo B was skewed and was compatible with the presence of two normally distributed subdistributions (mean values, 117 and 172 mg/dl). The analysis was repeated after stratification of individuals by low density lipoprotein (LDL) subclass phenotype. Among subjects with phenotype A (predominance of large, buoyant LDL), a single apo B distribution was found (mean, 115 mg/dl). Among subjects with phenotype B (predominance of small, dense LDL), the distribution was bimodal, with mean values, 116 and 167 mg/dl, similar to the unstratified data set. Thus the skewing of the overall apo B distribution in FCHL family members may be due to a distinct subset of individuals with phenotype B who are genetically susceptible to even higher elevations of apo B. The higher apo B/phenotype B subjects also showed significantly higher levels of triglyceride and LDL-cholesterol than the lower apo B/phenotype B subjects. The lower apo B/phenotype B subjects had higher triglyceride and lower LDL-cholesterol than the phenotype A subjects. The enhanced information regarding apo B and lipid levels in the three subgroups of individuals identified here may facilitate a better understanding of genetic susceptibility to coronary heart disease.