Mutual Information for Testing Gene-Environment Interaction

Despite current enthusiasm for investigation of gene-gene interactions and gene-environment interactions, the essential issue of how to define and detect gene-environment interactions remains unresolved. In this report, we define gene-environment interactions as a stochastic dependence in the context of the effects of the genetic and environmental risk factors on the cause of phenotypic variation among individuals. We use mutual information that is widely used in communication and complex system analysis to measure gene-environment interactions. We investigate how gene-environment interactions generate the large difference in the information measure of gene-environment interactions between the general population and a diseased population, which motives us to develop mutual information-based statistics for testing gene-environment interactions. We validated the null distribution and calculated the type 1 error rates for the mutual information-based statistics to test gene-environment interactions using extensive simulation studies. We found that the new test statistics were more powerful than the traditional logistic regression under several disease models. Finally, in order to further evaluate the performance of our new method, we applied the mutual information-based statistics to three real examples. Our results showed that P-values for the mutual information-based statistics were much smaller than that obtained by other approaches including logistic regression models.

The effect of genotype-by-environment interaction on longitudinal triglyceride profiles in the Bogalusa Heart Study

Triglyceride levels are a component of plasma lipids that are thought to be an important risk factor for coronary heart disease and are influenced by genetic and environmental factors, such as single nucleotide polymorphisms (SNPs), alcohol intake, and smoking. This study used longitudinal data from the Bogalusa Heart Study, a biracial community-based survey of cardiovascular disease risk factors. A sample of 1191 individuals, 4 to 38 years of age, was measured multiple times from 1973 to 2000. The study sample consisted of 730 white and 461 African American participants. Individual growth models were developed in order to assess gene-environment interactions affecting plasma triglycerides over time. After testing for inclusion of significant covariates and interactions, final models, each accounting for the effects of a different SNP, were assessed for fit and normality. After adjustment for all other covariates and interactions, LIPC -514C/T was found to interact with age3, age2, and age and a non-significant interaction of CETP -971G/A genotype with smoking status was found (p = 0.0812). Ever-smokers had higher triglyceride levels than never smokers, but persons heterozygous at this locus, about half of both races, had higher triglyceride levels after smoking cessation compared to current smokers. Since tobacco products increase free fatty acids circulating in the bloodstream, smoking cessation programs have the potential to ultimately reduce triglyceride levels for many persons. However, due to the effect of smoking cessation on the triglyceride levels of CETP -971G/A heterozygotes, the need for smoking prevention programs is also demonstrated. Both smoking cessation and prevention programs would have a great public health impact on minimizing triglyceride levels and ultimately reducing heart disease. ^

THE EFFECT OF PERSON-ENVIRONMENT INTERACTION ON ANXIETY, DEPRESSION, AND SUBJECTIVE DISTRESS OF OPTOMETRY STUDENTS

The purpose of this study was to investigate whether an incongruence between personality characteristics of individuals and concomitant charcteristics of health professional training environments on salient dimensions contributes to aspects of mental health. The dimensions examined were practical-theoretical orientation and the degree of structure-unstructure. They were selected for study as they are particularly important attributes of students and of learning environments. It was proposed that when the demand of the environment is disparate from the proclivities of the individual, strain arises. This strain was hypothesized to contribute to anxiety, depression, and subjective distress.^ Select subscales on the Omnibus Personality Inventory (OPI) were the operationalized measures for the personality component of the dimensions studied. An environmental index was developed to assess students' perceptions of the learning environment on these same dimensions. The Beck Depression Inventory, State-Trait Anxiety Inventory and General Well-Being schedule measured the outcome variables.^ A congruence model was employed to determine person-environment (P-E) interaction. Scores on the scales of the OPI and the environmental index were divided into high, medium, and low based on the range of scores. Congruence was defined as a match between the level of personality need and the complementary level of the perception of the environment. Alternatively, incongruence was defined as a mismatch between the person and the environment. The consistent category was compared to the inconsistent categories by an analysis of variance procedure. Furthermore, analyses of covariance were conducted with perceived supportiveness of the learning environment and life events external to the learning environment as the covariates. These factors were considered critical influences affecting the outcome measures.^ One hundred and eighty-five students (49% of the population) at the College of Optometry at the University of Houston participated in the study. Students in all four years of the program were equally represented in the study. However, the sample differed from the total population on representation by sex, marital status, and undergraduate major.^ The results of the study did not support the hypotheses. Further, after having adjusted for perceived supportiveness and life events external to the learning environment, there were no statistically significant differences between the congruent category and incongruent categories. Means indicated than the study sample experienced significantly lower depression and subjective distress than the normative samples.^ Results are interpreted in light of their utility for future study design in the investigation of the effects of P-E interaction. Emphasized is the question of the feasibility of testing a P-E interaction model with extant groups. Recommendations for subsequent research are proposed in light of the exploratory nature of the methodology. ^

BAYESIAN STATISTICAL METHODS IN GENE-ENVIRONMENT AND GENE-GENE INTERACTION STUDIES

Complex diseases such as cancer result from multiple genetic changes and environmental exposures. Due to the rapid development of genotyping and sequencing technologies, we are now able to more accurately assess causal effects of many genetic and environmental factors. Genome-wide association studies have been able to localize many causal genetic variants predisposing to certain diseases. However, these studies only explain a small portion of variations in the heritability of diseases. More advanced statistical models are urgently needed to identify and characterize some additional genetic and environmental factors and their interactions, which will enable us to better understand the causes of complex diseases. In the past decade, thanks to the increasing computational capabilities and novel statistical developments, Bayesian methods have been widely applied in the genetics/genomics researches and demonstrating superiority over some regular approaches in certain research areas. Gene-environment and gene-gene interaction studies are among the areas where Bayesian methods may fully exert its functionalities and advantages. This dissertation focuses on developing new Bayesian statistical methods for data analysis with complex gene-environment and gene-gene interactions, as well as extending some existing methods for gene-environment interactions to other related areas. It includes three sections: (1) Deriving the Bayesian variable selection framework for the hierarchical gene-environment and gene-gene interactions; (2) Developing the Bayesian Natural and Orthogonal Interaction (NOIA) models for gene-environment interactions; and (3) extending the applications of two Bayesian statistical methods which were developed for gene-environment interaction studies, to other related types of studies such as adaptive borrowing historical data. We propose a Bayesian hierarchical mixture model framework that allows us to investigate the genetic and environmental effects, gene by gene interactions (epistasis) and gene by environment interactions in the same model. It is well known that, in many practical situations, there exists a natural hierarchical structure between the main effects and interactions in the linear model. Here we propose a model that incorporates this hierarchical structure into the Bayesian mixture model, such that the irrelevant interaction effects can be removed more efficiently, resulting in more robust, parsimonious and powerful models. We evaluate both of the 'strong hierarchical' and 'weak hierarchical' models, which specify that both or one of the main effects between interacting factors must be present for the interactions to be included in the model. The extensive simulation results show that the proposed strong and weak hierarchical mixture models control the proportion of false positive discoveries and yield a powerful approach to identify the predisposing main effects and interactions in the studies with complex gene-environment and gene-gene interactions. We also compare these two models with the 'independent' model that does not impose this hierarchical constraint and observe their superior performances in most of the considered situations. The proposed models are implemented in the real data analysis of gene and environment interactions in the cases of lung cancer and cutaneous melanoma case-control studies. The Bayesian statistical models enjoy the properties of being allowed to incorporate useful prior information in the modeling process. Moreover, the Bayesian mixture model outperforms the multivariate logistic model in terms of the performances on the parameter estimation and variable selection in most cases. Our proposed models hold the hierarchical constraints, that further improve the Bayesian mixture model by reducing the proportion of false positive findings among the identified interactions and successfully identifying the reported associations. This is practically appealing for the study of investigating the causal factors from a moderate number of candidate genetic and environmental factors along with a relatively large number of interactions. The natural and orthogonal interaction (NOIA) models of genetic effects have previously been developed to provide an analysis framework, by which the estimates of effects for a quantitative trait are statistically orthogonal regardless of the existence of Hardy-Weinberg Equilibrium (HWE) within loci. Ma et al. (2012) recently developed a NOIA model for the gene-environment interaction studies and have shown the advantages of using the model for detecting the true main effects and interactions, compared with the usual functional model. In this project, we propose a novel Bayesian statistical model that combines the Bayesian hierarchical mixture model with the NOIA statistical model and the usual functional model. The proposed Bayesian NOIA model demonstrates more power at detecting the non-null effects with higher marginal posterior probabilities. Also, we review two Bayesian statistical models (Bayesian empirical shrinkage-type estimator and Bayesian model averaging), which were developed for the gene-environment interaction studies. Inspired by these Bayesian models, we develop two novel statistical methods that are able to handle the related problems such as borrowing data from historical studies. The proposed methods are analogous to the methods for the gene-environment interactions on behalf of the success on balancing the statistical efficiency and bias in a unified model. By extensive simulation studies, we compare the operating characteristics of the proposed models with the existing models including the hierarchical meta-analysis model. The results show that the proposed approaches adaptively borrow the historical data in a data-driven way. These novel models may have a broad range of statistical applications in both of genetic/genomic and clinical studies.

Gene by smoking interaction in blood pressure

Hypertension (HT) is mediated by the interaction of many genetic and environmental factors. Previous genome-wide linkage analysis studies have found many loci that show linkage to HT or blood pressure (BP) regulation, but the results were generally inconsistent. Gene by environment interaction is among the reasons that potentially explain these inconsistencies between studies. Here we investigate influences of gene by smoking (GxS) interaction on HT and BP in European American (EA), African American (AA) and Mexican American (MA) families from the GENOA study. A variance component-based method was utilized to perform genome-wide linkage analysis of systolic blood pressure (SBP), diastolic blood pressure (DBP), and HT status, as well as bivariate analysis for SBP and DBP for smokers, non-smokers, and combined groups. The most significant results were found for SBP in MA. The strongest signal was for chromosome 17q24 (LOD = 4.2), increased to (LOD = 4.7) in bivariate analysis but there was no evidence of GxS interaction at this locus (p = 0.48). Two signals were identified only in one group: on chromosome 15q26.2 (LOD = 3.37) in non-smokers and chromosome 7q21.11 (LOD = 1.4) in smokers, both of which had strong evidence for GxS interaction (p = 0.00039 and 0.009 respectively). There were also two other signals, one on chromosome 20q12 (LOD = 2.45) in smokers, which became much higher in the combined sample (LOD = 3.53), and one on chromosome 6p22.2 (LOD = 2.06) in non-smokers. Neither peak had very strong evidence for GxS interaction (p = 0.08 and 0.06 respectively). A fine mapping association study was performed using 200 SNPs in 30 genes located under the linkage signals on chromosomes 15 and 17. Under the chromosome 15 peak, the association analysis identified 6 SNPs accounting for a 7 mmHg increase in SBP in MA non-smokers. For the chromosome 17 linkage peak, the association analysis identified 3 SNPs accounting for a 6 mmHg increase in SBP in MA. However, none of these SNPs was significant after correcting for multiple testing, and accounting for them in the linkage analysis produced very small reductions in the linkage signal. ^ The linkage analysis of BP traits considering the smoking status produced very interesting signals for SBP in the MA population. The fine mapping association analysis gave some insight into the contribution of some SNPs to two of the identified signals, but since these SNPs did not remain significant after multiple testing correction and did not explain the linkage peaks, more work is needed to confirm these exploratory results and identify the culprit variations under these linkage peaks. ^

Genetics and epidemiology of gallbladder disease in New World native peoples.

Native peoples of the New World, including Amerindians and admixed Latin Americans such as Mexican-Americans, are highly susceptible to diseases of the gallbladder. These include cholesterol cholelithiasis (gallstones) and its complications, as well as cancer of the gallbladder. Although there is clearly some necessary dietary or other environmental risk factor involved, the pattern of disease prevalence is geographically associated with the distribution of genes of aboriginal Amerindian origin, and levels of risk generally correspond to the degree of Amerindian admixture. This pattern differs from that generally associated with Westernization, which suggests a gene-environment interaction, and that within an admixed population there is a subset whose risk is underestimated when admixture is ignored. The risk that an individual of a susceptible New World genotype will undergo a cholecystectomy by age 85 can approach 40% in Mexican-American females, and their risk of gallbladder cancer can reach several percent. These are heretofore unrecognized levels of risk, especially of the latter, because previous studies have not accounted for admixture or for the loss of at-risk individuals due to cholecystectomy. A genetic susceptibility may, thus, be as "carcinogenic" in New World peoples as any known major environmental exposure; yet, while the risk has a genetic basis, its expression as gallbladder cancer is so delayed as to lead only very rarely to multiply-affected families. Estimates in this paper are derived in part from two studies of Mexican-Americans in Starr County and Laredo, Texas.

Developmental programming of tumor suppressor gene penetrance

Epidemiological studies have led to the hypothesis that major risk factors for developing diseases such as hypertension, cardiovascular disease and adult-onset diabetes are established during development. This developmental programming hypothesis proposes that exposure to an adverse stimulus or insult at critical, sensitive periods of development can induce permanent alterations in normal physiological processes that lead to increased disease risk later in life. For cancer, inheritance of a tumor suppressor gene defect confers a high relative risk for disease development. However, these defects are rarely 100% penetrant. Traditionally, gene-environment interactions are thought to contribute to the penetrance of tumor suppressor gene defects by facilitating or inhibiting the acquisition of additional somatic mutations required for tumorigenesis. The studies presented herein identify developmental programming as a distinctive type of gene-environment interaction that can enhance the penetrance of a tumor suppressor gene defect in adult life. Using rats predisposed to uterine leiomyoma due to a germ-line defect in one allele of the tuberous sclerosis complex 2 (Tsc-2) tumor suppressor gene, these studies show that early-life exposure to the xenoestrogen, diethylstilbestrol (DES), during development of the uterus increased tumor incidence, multiplicity and size in genetically predisposed animals, but failed to induce tumors in wild-type rats. Uterine leiomyomas are ovarian-hormone dependent tumors that develop from the uterine myometrium. DES exposure was shown to developmentally program the myometrium, causing increased expression of estrogen-responsive genes prior to the onset of tumors. Loss of function of the normal Tsc-2 allele remained the rate-limiting event for tumorigenesis; however, tumors that developed in exposed animals displayed an enhanced proliferative response to ovarian steroid hormones relative to tumors that developed in unexposed animals. Furthermore, the studies presented herein identify developmental periods during which target tissues are maximally susceptible to developmental programming. These data suggest that exposure to environmental factors during critical periods of development can permanently alter normal physiological tissue responses and thus lead to increased disease risk in genetically susceptible individuals. ^

Modifier genes and susceptibility to colorectal cancer in individuals with Lynch syndrome

Lynch syndrome, is caused by inherited germ-line mutations in the DNA mismatch repair genes resulting in cancers at an early age, predominantly colorectal (CRC) and endometrial cancers. Though the median age at onset for CRC is about 45 years, disease penetrance varies suggesting that cancer susceptibility may be modified by environmental or other low-penetrance genes. Genetic variation due to polymorphisms in genes encoding metabolic enzymes can influence carcinogenesis by alterations in the expression and activity level of the enzymes. Variation in MTHFR, an important folate metabolizing enzyme can affect DNA methylation and DNA synthesis and variation in xenobiotic-metabolizing enzymes can affect the metabolism and clearance of carcinogens, thus modifying cancer risk. ^ This study examined a retrospective cohort of 257 individuals with Lynch syndrome, for polymorphisms in genes encoding xenobiotic-metabolizing enzymes-- CYP1A1 (I462V and MspI), EPHX1 (H139R and Y113H), GSTP1 (I105V and A114V), GSTM1 and GSTT1 (deletions) and folate metabolizing enzyme--MTHFR (C677T and A1298C). In addition, a series of 786 cases of sporadic CRC were genotyped for CYP1A1 I462V and EPHX1 Y113H to assess gene-gene interaction and gene-environment interaction with smoking in a case-only analysis. ^ Prominent findings of this study were that the presence of an MTHFR C677T variant allele was associated with a 4 year later age at onset for CRC on average and a reduced age-associated risk for developing CRC (Hazard ratio: 0.55; 95% confidence interval: 0.36–0.85) compared to the absence of any variant allele in individuals with Lynch syndrome. Similarly, Lynch syndrome individuals heterozygous for CYP1A1 I462V A>G polymorphism developed CRC an average of 4 years earlier and were at a 78% increased age-associated risk (Hazard ratio for AG relative to AA: 1.78; 95% confidence interval: 1.16-2.74) than those with the homozygous wild-type genotype. Therefore these two polymorphisms may be additional susceptibility factors for CRC in Lynch syndrome. In the case-only analysis, evidence of gene-gene interaction was seen between CYP1A1 I462V and EPHX1 Y113H and between EPHX1 Y113H and smoking suggesting that genetic and environmental factors may interact to increase sporadic CRC risk. Implications of these findings are the ability to identify subsets of high-risk individuals for targeted prevention and intervention. ^