A semiparametric isotonic regression model for skewed distributions with application to DNA-RNA-protein analysis.

In this paper, we propose a semiparametric regression model that is built upon an isotonic regression model with the assumption that the random error follows a skewed distribution. We develop an expectation-maximization algorithm for obtaining the maximum likelihood estimates of the model parameters, examine the asymptotic properties of the estimators, conduct simulation studies to explore the performance of the proposed model, and apply the method to evaluate the DNA-RNA-protein relationship and identify genes that are key factors in tumor progression.

Pool adjacent violators algorithm-assisted learning with application on estimating optimal individualized treatment regimes.

Personalized medicine allows individuals to choose the best fit of their treatments based on their characteristics through an individualized treatment regime. In this paper, we develop a pool adjacent violators algorithm-assisted learning method to find the optimal individualized treatment regime under the monotone single-index outcome gain model. The proposed estimator is more efficient than peers, and it is robust to the misspecification of the propensity score model or the baseline regression model. The optimal treatment regime is also robust to the misspecification of the functional form of the expected outcome gain model. Simulation studies verified our theoretical results. We also provide an estimate of the expected outcome gain model. Plotting the expected outcome gain versus an individual's characteristics index can visualize how significant the treatment effect is over the control. We apply the proposed method to an AIDS study.

Set-regression with applications to subgroup analysis.

Regression is a commonly used statistical model. It is the conditional mean of the response given covariates µ(x)=E(Y|X=x) . However, in some practical problems, the interest is the conditional mean of the response given the covariates belonging to some set A. Notably, in precision medicine and subgroup analysis in clinical trials, the aim is to identify subjects who benefit the most from the treatment, or identify an optimal set in the covariate space which manifests treatment favoritism if a subject's covariates fall in this set and the subject is classified to the favorable treatment subgroup. Existing methods for subgroup analysis achieve this indirectly by using classical regression. This motivates us to develop a new type of regression: set-regression, defined as µ(A)=E(Y|X∈A) which directly addresses the subgroup analysis problem. This extends not only the classical regression model but also improves recursive partitioning and support vector machine approaches, and is particularly suitable for objectives involving optimization of the regression over sets, such as subgroup analysis. We show that the new versatile set-regression identifies the subgroup with increased accuracy. It is easy to use. Simulation studies also show superior performance of the proposed method in finite samples.

Robust estimates of regional treatment effects in multiregional randomized clinical trials with ordinal responses.

Global clinical trials involving multiple regions are common in current drug development processes. Determining the regional treatment effects of a new therapy over an existing therapy is important to both the sponsors and the regulatory agencies in the regions. Existing methods are mainly for continuous primary endpoints and use subjectively specified models, which may deviate from the true model. Here, we consider trials that have ordinal responses as the primary endpoint. This article extends the recently developed robust semiparametric ordinal regression model to estimate regional treatment effects, in which the regression coefficients and regional effects are modeled parametrically for ease of interpretation, and the regression link function is specified nonparametrically for robustness. The model parameters are estimated by semiparametric maximum likelihood estimation, and the null hypothesis of no regional effect is tested by the Wald test. Simulation studies are conducted to evaluate the performance of the proposed method and compare it with the commonly used parametric model. The results of the former show an improved overall performance over the latter. In particular, the model yields much higher precision in estimation and prediction than the fixed-link model. This result is especially appealing since our interest is to estimate the treatment effect more efficiently and the estimand is of particular interest in multiregional clinical trials. We then apply the method by analyzing real multiregional clinical trials with ordinal responses as their primary endpoint.

Robust estimates of regional treatment effects in multiregional randomized clinical trials with semiparametric logistic model.

In multiregional randomized clinical trials (MRCTs), determining the regional treatment effect of a new treatment over an existing one is important to both the sponsor and related regulatory agencies. Also of particular interest is to test the null hypothesis that the treatment benefit is the same among all the regions. Existing methods are mainly for continuous endpoint and use parametric models, which are not robust. MRCTs are known for facing increased variation and heterogeneity and a robust model for its design and analysis would be desirable. We consider clinical trials with a binary primary endpoint and propose a robust semiparametric logistic model which has a known parametric and an unknown nonparametric component. The parametric component represents our prior knowledge about the model, and the nonparametric part reflects uncertainty. Compared to the classic logistic model for this problem, the proposed model has the following advantages: robust to model assumption, more flexible and accurate to model the relationship between the response and covariates, and possibly more accurate parameter estimates. The model parameters are estimated by profile maximum likelihood approach, and the null hypothesis of regional treatment difference being the same is tested by the profile likelihood ratio statistic. Asymptotic properties of the estimates are derived. Simulation studies are conducted to evaluate the performance of the proposed model, which demonstrated clear advantages over the classic logistic model. The method is then applied to analyzing a real MRCT.

Identification of subgroups via partial linear regression modeling approach.

In clinical trials, treatment effects often vary from subject to subject. Some subjects may benefit more than others from a specific treatment. One of the aims of subgroup analysis is to identify if there are subgroups of subjects with differential treatment effects. As in standard analysis, we first test if subgroups with differential treatment effects exist; if they do, we classify the subjects into different subgroups based on their covariate profiles; otherwise, we conclude no subgroups have differential treatment effects in this population. Existing methods utilize regression models, particularly linear models, for such analysis. However, in practice, not all effects of covariates on responses are linear. To address this issue, the article proposes a more flexible model, the partial linear model with a nonlinear monotone function to describe some specific effects of covariates and with a linear component to describe the effects of other covariates, develops model-fitting algorithm and derives model asymptotics. We then utilize the Wald statistic to test the existence of subgroups and the Neyman-Pearson rule to classify subjects into the subgroups. Simulation studies are conducted to evaluate the finite sample performance of the proposed method by comparing it with the commonly used linear models. Finally, we apply the methods to analyzing a real clinical trial.

Semiparametric isotonic regression modelling and estimation for group testing data.

In the group testing procedure, several individual samples are grouped and the pooled samples, instead of each individual sample, are tested for outcome status (e.g., infectious disease status). Although this cost-effectiveness strategy in data collection is both labor and time efficient, it poses statistical challenges to derive statistically and computationally efficient estimators under semiparametric models. We consider semiparametric isotonic regression models for the simultaneous estimation of the conditional probability curve and covariate effects, in which a parametric form for combining the covariate information is assumed and the monotonic link function is left unspecified. We develop an expectation-maximization algorithm to overcome the computational challenge and embed the pool-adjacent violators algorithm in the M-step to facilitate the computation. We establish the large sample behavior of the proposed estimators and examine their finite sample performance in simulation studies. We apply the proposed method to data from the National Health and Nutrition Examination Survey for illustration.

Linear mixed models for association analysis of quantitative traits with next-generation sequencing data.

We develop linear mixed models (LMMs) and functional linear mixed models (FLMMs) for gene-based tests of association between a quantitative trait and genetic variants on pedigrees. The effects of a major gene are modeled as a fixed effect, the contributions of polygenes are modeled as a random effect, and the correlations of pedigree members are modeled via inbreeding/kinship coefficients. F -statistics and χ 2 likelihood ratio test (LRT) statistics based on the LMMs and FLMMs are constructed to test for association. We show empirically that the F -distributed statistics provide a good control of the type I error rate. The F -test statistics of the LMMs have similar or higher power than the FLMMs, kernel-based famSKAT (family-based sequence kernel association test), and burden test famBT (family-based burden test). The F -statistics of the FLMMs perform well when analyzing a combination of rare and common variants. For small samples, the LRT statistics of the FLMMs control the type I error rate well at the nominal levels α = 0.01 and 0.05 . For moderate/large samples, the LRT statistics of the FLMMs control the type I error rates well. The LRT statistics of the LMMs can lead to inflated type I error rates. The proposed models are useful in whole genome and whole exome association studies of complex traits.

Measuring gene-gene interaction using Kullback-Leibler divergence.

Genome-wide association studies (GWAS) are used to investigate genetic variants contributing to complex traits. Despite discovering many loci, a large proportion of "missing" heritability remains unexplained. Gene-gene interactions may help explain some of this gap. Traditionally, gene-gene interactions have been evaluated using parametric statistical methods such as linear and logistic regression, with multifactor dimensionality reduction (MDR) used to address sparseness of data in high dimensions. We propose a method for the analysis of gene-gene interactions across independent single-nucleotide polymorphisms (SNPs) in two genes. Typical methods for this problem use statistics based on an asymptotic chi-squared mixture distribution, which is not easy to use. Here, we propose a Kullback-Leibler-type statistic, which follows an asymptotic, positive, normal distribution under the null hypothesis of no relationship between SNPs in the two genes, and normally distributed under the alternative hypothesis. The performance of the proposed method is evaluated by simulation studies, which show promising results. The method is also used to analyze real data and identifies gene-gene interactions among RAB3A, MADD, and PTPRN on type 2 diabetes (T2D) status.

A class of semiparametric cure models with current status data.

Current status data occur in many biomedical studies where we only know whether the event of interest occurs before or after a particular time point. In practice, some subjects may never experience the event of interest, i.e., a certain fraction of the population is cured or is not susceptible to the event of interest. We consider a class of semiparametric transformation cure models for current status data with a survival fraction. This class includes both the proportional hazards and the proportional odds cure models as two special cases. We develop efficient likelihood-based estimation and inference procedures. We show that the maximum likelihood estimators for the regression coefficients are consistent, asymptotically normal, and asymptotically efficient. Simulation studies demonstrate that the proposed methods perform well in finite samples. For illustration, we provide an application of the models to a study on the calcification of the hydrogel intraocular lenses.

Subgroup analysis with semiparametric models toward precision medicine.

In analyzing clinical trials, one important objective is to classify the patients into treatment-favorable and nonfavorable subgroups. Existing parametric methods are not robust, and the commonly used classification rules ignore the fact that the implications of treatment-favorable and nonfavorable subgroups can be different. To address these issues, we propose a semiparametric model, incorporating both our knowledge and uncertainty about the true model. The Wald statistics is used to test the existence of subgroups, while the Neyman-Pearson rule to classify each subject. Asymptotic properties are derived, simulation studies are conducted to evaluate the performance of the method, and then method is used to analyze a real-world trial data.

Evaluating accuracy of diagnostic tests without conditional independence assumption.

Evaluating the accuracy (ie, estimating the sensitivity and specificity) of new diagnostic tests without the presence of a gold standard is of practical meaning and has been the subject of intensive study for several decades. Existing methods use 2 or more diagnostic tests under several basic assumptions and then estimate the accuracy parameters via the maximum likelihood estimation. One of the basic assumptions is the conditional independence of the tests given the disease status. This assumption is impractical in many real applications in veterinary research. Several methods have been proposed with various dependence models to relax this assumption. However, these methods impose subjective dependence structures, which may not be practical and may introduce additional nuisance parameters. In this article, we propose a simple method for addressing this problem without the conditional independence assumption, using an empirical conditioning approach. The proposed method reduces to the popular Hui-Walter model in the case of conditional independence. Also, our likelihood function is of order-2 polynomial in parameters, while that of Hui-Walter is of order-3. The reduced model complexity increases the stability in estimation. Simulation studies are conducted to evaluate the performance of the proposed method, which shows overall smaller biases in estimation and is more stable than the existing method, especially when tests are conditionally dependent. Two real data examples are used to illustrate the proposed method.

Genetic risks and genetic model specification.

Genetic risks and genetic models are often used in design and analysis of genetic epidemiology studies. A genetic model is defined in terms of two genetic risk measures: genotype relative risk and odds ratio. The impacts of choosing a risk measure on the resulting genetic models are studied in the power to detect association and deviation from Hardy-Weinberg equilibrium in cases using genetic relative risk. Extensive simulations demonstrate that the power of a study to detect associations using odds ratio is lower than that using relative risk with the same value when other parameters are fixed. When the Hardy-Weinberg equilibrium holds in the general population, the genetic model can be inferred by the deviation from Hardy-Weinberg equilibrium in only cases. Furthermore, it is more efficient than that based on the deviation from Hardy-Weinberg equilibrium in all cases and controls.

Rank-based tests for identifying multiple genetic variants associated with quantitative traits.

We consider the analysis of multiple genetic variants within a gene or a region that are expected to confer risks to human complex diseases with quantitative traits, where the trait values do not follow the normal distribution even after some transformations. We rank the phenotypic values, calculate a score to measure the trend effect of a particular allele for each marker, and then construct three statistics based on the quadratic frameworks of methods Hotelling T(2) , the summation of squared univariate statistic and the inverse of the square root weighted statistics to combine the scores for different marker loci. Simulation results show that the above three test statistics can control the type I error rate well and are more robust than standard tests constructed based on linear regression. Application to GAW16 data for rheumatoid arthritis successfully detects the association between the HLA-DRB1 gene and anticyclic citrullinated protein measure, while the standard methods based on normal assumption cannot detect this association.

Semiparametric partial linear modeling of risk factors for ear infections: the Early Childhood Longitudinal Study.

The Early Childhood Longitudinal Study-Kindergarten Class of 2010-2011 (ECLS-K:2011) ascertained timing of ear infections within age specified intervals and parent's/caregiver's report of medically diagnosed hearing loss. In this nationally representative, school-based sample of children followed from kindergarten entry through fifth grade, academic performance in reading, mathematics, and science was assessed longitudinally. Prior investigations of this ECLS-K:2011 cohort showed that age has a non-linear, monotonically increasing functional relationship with academic performance. Because of this knowledge, a semiparametric partial linear model is proposed, in which the effect of age is modeled by an unknown monotonically increasing function along with other regression parameters. The parameters are estimated by a semiparametric maximum likelihood estimator. A test of a constant effect of age is also proposed. Simulation studies are conducted to evaluate the performance of the proposed method, as compared with the commonly used linear model; the former outperforms the latter based on several criteria. We then analyzed ECLS-K:2011 data to compare results of the partial linear parametric model estimation with that of classical linear regression models.

Impact on modes of inheritance and relative risks of using extreme sampling when designing genetic association studies.

Using extreme phenotypes for association studies can improve statistical power . We study the impact of using samples with extremely high or low traits on the alternative model space, the genotype relative risks, and the genetic models in association studies. We prove the following results: when the risk allele causes high-trait values, the more extreme the high traits, the larger the genotype relative risks, which is not always true for using extreme low traits; we also prove that a genetic model theoretically changes with more extreme trait except for the recessive or dominant models. Practically, however, the impact of deviations from the true genetic model at a functional locus due to selective sampling is virtually negligible. The implications of our findings are discussed. Numerical values are reported for illustrations.

Case-Control Genome-wide Joint Association Study Using Semiparametric Empirical Model and Approximate Bayes Factor.

We propose a semiparametric approach for the analysis of case-control genome-wide association study. Parametric components are used to model both the conditional distribution of the case status given the covariates and the distribution of genotype counts, whereas the distribution of the covariates are modeled nonparametrically. This yields a direct and joint modeling of the case status, covariates and genotype counts, and gives better understanding of the disease mechanism and results in more reliable conclusions. Side information, such as the disease prevalence, can be conveniently incorporated into the model by empirical likelihood approach and leads to more efficient estimates and powerful test in the detection of disease-associated SNPs. Profiling is used to eliminate a nuisance nonparametric component, and the resulting profile empirical likelihood estimates are shown to be consistent and asymptotically normal. For the hypothesis test on disease association, we apply the approximate Bayes factor (ABF) which is computationally simple and most desirable in genome-wide association studies where hundreds of thousands to a million genetic markers are tested. We treat the approximate Bayes factor as a hybrid Bayes factor which replaces the full data by the maximum likelihood estimates of the parameters of interest in the full model and derive it under a general setting. The deviation from Hardy-Weinberg Equilibrium (HWE) is also taken into account and the ABF for HWE using cases is shown to provide evidence of association between a disease and a genetic marker. Simulation studies and an application are further provided to illustrate the utility of the proposed methodology.

Maximum Likelihood Estimation for Shape-restricted Single-index Hazard Models.

Single-index models are becoming increasingly popular in many scientific applications as they offer the advantages of flexibility in regression modeling as well as interpretable covariate effects. In the context of survival analysis, the single-index hazards models are natural extensions of the Cox proportional hazards models. In this paper, we propose a novel estimation procedure for single-index hazard models under a monotone constraint of the index. We apply the profile likelihood method to obtain the semiparametric maximum likelihood estimator, where the novelty of the estimation procedure lies in estimating the unknown monotone link function by embedding the problem in isotonic regression with exponentially distributed random variables. The consistency of the proposed semiparametric maximum likelihood estimator is established under suitable regularity conditions. Numerical simulations are conducted to examine the finite-sample performance of the proposed method. An analysis of breast cancer data is presented for illustration.

Bayes factor based on the trend test incorporating Hardy-Weinberg disequilibrium: more power to detect genetic association.

In the analysis of case-control genetic association, the trend test and Pearson's test are the two most commonly used tests. In genome-wide association studies (GWAS), Bayes factor (BF) is a useful tool to support significant P-values, and a better measure than P-value when results are compared across studies with different sample sizes. When reporting the P-value of the trend test, we propose a BF directly based on the trend test. To improve the power to detect association under recessive or dominant genetic models, we propose a BF based on the trend test and incorporating Hardy-Weinberg disequilibrium in cases. When the true model is unknown, or both the trend test and Pearson's test or other robust tests are applied in genome-wide scans, we propose a joint BF, combining the previous two BFs. All three BFs studied in this paper have closed forms and are easy to compute without integrations, so they can be reported along with P-values, especially in GWAS. We discuss how to use each of them and how to specify priors. Simulation studies and applications to three GWAS are provided to illustrate their usefulness to detect nonadditive gene susceptibility in practice.

Detecting case-control expression quantitative trait loci using locally most powerful or maximin robust rank tests.

In testing genome-wide gene expression quantitative trait loci, efficiency robust statistical methods and their computational convenience are most relevant. For this purpose, we propose to use a modified locally most powerful rank test for the analysis of case-control expression data. This modified rank test statistic is computationally simple, robust for non-normally distributed expression data, and asymptotically locally most powerful. It depends on the specification of a location distribution form for data but is not sensitive to misspecifications. When such a location distribution form cannot be specified, we apply Gastwirth's maximin efficiency robust rank test to gene expression data to maximize the worst Pitman asymptotic relative efficiency among a family of location distributions. We conduct simulation studies to assess their performance and use an application to real data for illustration.