A Dirichlet process mixture model for clustering longitudinal gene expression data

Subgroup identification (clustering) is an important problem in biomedical research. Gene expression profiles are commonly utilized to define subgroups. Longitudinal gene expression profiles might provide additional information on disease progression than what is captured by baseline profiles alone. Therefore, subgroup identification could be more accurate and effective with the aid of longitudinal gene expression data. However, existing statistical methods are unable to fully utilize these data for patient clustering. In this article, we introduce a novel clustering method in the Bayesian setting based on longitudinal gene expression profiles. This method, called BClustLonG, adopts a linear mixed‐effects framework to model the trajectory of genes over time, while clustering is jointly conducted based on the regression coefficients obtained from all genes. In order to account for the correlations among genes and alleviate the high dimensionality challenges, we adopt a factor analysis model for the regression coefficients. The Dirichlet process prior distribution is utilized for the means of the regression coefficients to induce clustering. Through extensive simulation studies, we show that BClustLonG has improved performance over other clustering methods. When applied to a dataset of severely injured (burn or trauma) patients, our model is able to identify interesting subgroups. Copyright © 2017 John Wiley & Sons, Ltd.     

Socio-demographic patient profiles and hospital efficiency: does patient mix affect a hospital's ability to perform?

This study investigates whether inclusion of patient profiles impacts on the inferences drawn from measuring performance using patient level data. Performance is in this setting defined by resources used in treating patients in a given diagnose related group where use of resources is approximated by length of stay (LOS). The analysis is based on Danish registry data from 2006. Patient data include registry data on income, employment status and information on whether the patient receives benefits or lives alone. Considerable variation in the socio-demographic characteristics of patients across Danish hospitals was observed, and some patient characteristics were shown to drive the need for longer hospital stays beyond what is captured in DRG scores. Ranking of hospitals based on observed versus expected LOS remained largely unaffected when controlling for patient characteristics, suggesting that variation in LOS across hospitals is mainly driven by other factors than patients’ socio-demographic characteristics. Nevertheless, the results of this study indicate that the current Danish remuneration system discriminates hospitals that more often serve older patients and patients with a less developed social network. These hospitals tend to have a reduced turnover of patients and their ability to generate revenue is therefore constrained.     

Bayesian models for population-based case-control studies when the population is in Hardy-Weinberg equilibrium

Association analysis of genetic polymorphisms has been mostly performed in a case-control setting with unrelated affected subjects compared with unrelated unaffected subjects. In this paper, we present a Bayesian method for analyzing such case-control data when the population is in Hardy-Weinberg equilibrium. Our Bayesian method depends on the informative prior which is the retrospective likelihood based on historical data, raised to a power a. By modeling the retrospective likelihood properly, different prior information about the studied population can be incorporated into the specification of the prior. The scalar a is a precision parameter quantifying the heterogeneity between current and historical data. A guide value for a is discussed in this paper. The informative prior and posterior distributions are proper under very general conditions. Therefore, our method can be applied in most case-control studies. Further, for assessing gene-environment interactions, our approach will naturally lead to a Bayesian model depending only on the case data, when genotype and environmental factors are independent in the population. Thus our approach can be applied to case-only studies. A real example is used to show the applications of our method.     

Commentary: meta-analysis of individual participants' data in genetic epidemiology

The authors summarize their experience in the conduct of meta-analysis of individual participants' data (MIPD) with time-to-event analyses in the field of genetic epidemiology. The MIPD offers many advantages compared with a meta-analysis of the published literature. These include standardization of case definitions, outcomes, and covariates; inclusion of updated information; the ability to fully test the assumptions of time-to-event models; better control of confounding; standardization of analyses of genetic loci that are in linkage disequilibrium; evaluation of alternative genetic models and multiple genes; consistent treatment of subpopulations; assessment of sampling bias; and the establishment of an international collaboration with the capability to prospectively update the meta-analyses and synthesize new information on multiple genetic loci and outcomes. The disadvantages of a MIPD compared with a meta-analysis of the published literature are that a much greater commitment of time and resources is required to collect primary data and to coordinate a large collaborative project. An MIPD may collect additional, unpublished data, but it is possible that not all published data may be contributed at the individual level. For questions that justify the required intensive effort, the MIPD method is a useful tool to help clarify the role of candidate genes in complex human diseases.     

A score test for genetic class‐level association with nonlinear biomarker trajectories

Emerging data suggest that the genetic regulation of the biological response to inflammatory stress may be fundamentally different to the genetic underpinning of the homeostatic control (resting state) of the same biological measures. In this paper, we interrogate this hypothesis using a single‐SNP score test and a novel class‐level testing strategy to characterize protein‐coding gene and regulatory element‐level associations with longitudinal biomarker trajectories in response to stimulus. Using the proposed c lass‐level a ssociation s core s tatistic for l ongitudinal d ata, which accounts for correlations induced by linkage disequilibrium, the genetic underpinnings of evoked dynamic changes in repeatedly measured biomarkers are investigated. The proposed method is applied to data on two biomarkers arising from the Genetics of Evoked Responses to Niacin and Endotoxemia study, a National Institutes of Health‐sponsored investigation of the genomics of inflammatory and metabolic responses during low‐grade endotoxemia. Our results suggest that the genetic basis of evoked inflammatory response is different than the genetic contributors to resting state, and several potentially novel loci are identified. A simulation study demonstrates appropriate control of type‐1 error rates, relative computational efficiency, and power. Copyright © 2017 John Wiley & Sons, Ltd.     

Priority setting in health authorities: a novel approach to a historical activity

As resources in health care are scarce, health authorities and other health organizations are charged with determining how best to spend limited resources. While a number of formal approaches to priority setting within health authorities have been used internationally, there has been limited success with such activity, particularly across major service portfolios. This participatory action research project instituted a novel priority setting framework, coined macro-marginal analysis (MMA), in a fully integrated urban health region in Alberta, Canada. The focus of MMA is on identifying areas for service growth and areas for resource release, then determining, based on pre-defined, locally generated criteria, if actual shifts or re-allocation of resources should occur. For fiscal year 2002/03, the Calgary Health Region identified over 40 M dollars in resource releases (approximately 3% of the total budget), which were made available for servicing the deficit, and more importantly for our purposes, re-investing in service growth areas. The MMA framework is pragmatic in nature and has the ability to incorporate relevant evidence directly into the decision-making process. This work constitutes a significant advancement in health economics, and responds where previous priority setting approaches have failed in that it allows decision-makers to achieve genuine re-allocation of resources with the aim of improving population health or better meeting other important criteria.     

Monitoring human radiation exposure by gene expression profiling: possibilities and pitfalls

Advances in high throughput analysis of mRNA expression have made it possible to establish gene expression profiles for different cells, tissues, diseases and exposure states. For instance, recent studies have demonstrated the utility of such an approach to classify sub-types of cancers with more detail than was previously possible. In addition, gene expression studies of ionizing radiation exposure both in vitro and in vivo are affording insight into the molecular mechanisms of mammalian radiation response. We have demonstrated that radiation expression profiles are a good predictor of p53 function in cell lines, and such profiles also indicate a major role for p53-regulated genes in the in vivo radiation response. Gene expression can be a sensitive indicator of radiation response as we have shown linear dose-responses for induction of several genes down to doses as low as 2 cGy. As profiles are established from radiation studies, it is hoped that they may be useful for identifying individuals with specific exposures or predisposition to negative outcome of exposure. Although this technology holds great promise, some obstacles remain to be overcome before it can be successfully applied to population studies.     

DNA chips for gene expression analysis and their application in diagnostics

Due to large genome projects more and more genetic information and more and more data about genetic background of various diseases can be obtained. By means of microarrays thousands to hundreds of thousands of molecular analysis can be performed at a time, rapidly, under identical conditions, hereby microarrays provide a solution to continuously increasing throughput demands from large genomic, pharmacogenetic and proteomic programmes. Microarrays offer unprecedented opportunities for qualitative and quantitative analysis of gene expression, understanding gene function and detection gene alterations, polimorphisms. Microarray expression profiles will define stages of different tumor types and new prognostic subgroups in cancer and other diseases. For individual patients, more precise diagnosis and risk assessment based on expression profiles are already achievable, leading to more accurate determination of prognosis and more individually tailored treatment. Microarrays will play an important role in diagnostics of infectious diseases and identification of novel diagnostic and therapeutic targets. Wide-spread application of microarrays in clinical practice is not typical today, but their production on an industrial scale and the increased information provided by microarrays should assure their entry into clinical practice as very powerful, sensitive, specific and rapid tools of diagnostics of various diseases.     

数字化医院大型医疗设备绩效评价方式的改进研究

目的：数字化医院是通过宽带网络把数字化医疗设备、数字化医学影像系统和数字化医疗信息系统等全部临床作业过程纳入到数字化网络中,通过大型医疗设备的绩效评估监控管理设备的使用。方法：对大型医疗设备进行有效的绩效评估,可以促进其资源配置的优化、精准科学的管理,合理良性的应用,使大型医疗设备充分有效地应用到临床诊疗中。结果：基础数据的搜集,是大型医疗设备绩效评估的基础。基础数据的即时有效、真实准确以及规范科学可以增强大型医疗设备综合绩效管理评估的可操作性与科学性。结论：本研究通过设定大型医疗设备使用的基础数据采集规则,从而提高了大型医疗设备统计中基础数据的时效性、准确性和规范性。     

Phylogenetic character mapping of RADES Probing,a new marker for exploring the clonal evolution of expressed coding sequences in Trypanosoma cruzi,the agent of Chagas disease

We have tested a new genetic marker, RADES Probing (RADES-P), on a standard sample of 19 laboratory-cloned stocks of Trypanosoma cruzi, the agent of Chagas disease. This set of stocks, fully characterized using multilocus enzyme electrophoresis (MLEE) and random amplified polymorphic DNA (RAPD), is representative of this parasite’s main genetic subdivisions. RADES-P consists in hybridizing RAPD profiles with probes composed of the products of random amplified differentially expressed sequences (RADES). The profiles thus obtained uncover only expressed coding sequences that are as well present on RAPD gels. Direct visual examination and the banding record show that these RADES-P profiles are different of, and not redundant with, both RAPD and RADES patterns obtained on the same set of stocks with the same primers. Phylogenetic character mapping (PCM) of the RADES-P polymorphism fairly confirms the known population structure and phylogenetic diversity of T. cruzi. This suggests that the impact of clonal evolution on T. cruzi has been predominant enough over the long term to carve the polymorphism of all types of DNA sequences, including polymorphisms of expressed coding sequences, although these sequences are subject to natural selection.     

The pursuit of equity in NHS resource allocation: should morbidity replace utilisation as the basis for setting health care capitations?

Although the English NHS has been described as a world leader in pioneering methods of distributing expenditure in relation to population needs, concerns about the legitimacy of using the current utilisation-based model to allocate health service resources are mounting. In this paper, we present a critical review of NHS resource allocation in England and demonstrate the feasibility and impact of using direct health estimates as a basis for setting health care capitations. Comparing target allocations for the inpatient treatment of coronary heart disease in a sample of 34 primary care trusts in contrasting locations in England, we find that a morbidity-based model would result in a significant shift in hospital resources away from deprived areas, towards areas with older demographic profiles and towards rural areas. Discussing the findings in relation to a wider policy context that is generally concerned to direct more health care resources towards the poor, the paper concludes by calling for greater clarity between the goals of health care equity and health equity. Whilst the former demands that the legitimate needs of demographically older populations for more health care resources are acknowledged, the goal of health equity requires real political commitment to resource broader social policy initiatives.     

Genetic epidemiology: a review of the statistical basis

Genetic epidemiology is the analysis of the familial distributions of traits, with a view to understanding any possible genetic basis; the fundamentals of this area are reviewed. The terminology of genetic models is fully described, since it is this wealth of terminology which can make it difficult for a statistician to enter the area. The statistical approaches to the analysis of data on sets of interrelated individuals are then discussed, and resulting problems of sampling design and analysis are considered.     

Nested, non-parametric, correlative analysis of microarrays for heterogenous phenotype characterization

We present a non-parametric approach for qualitatively selecting candidate genes to characterize several criteria that are nested among genes selected on the basis of their individual, similar effects upon an array-wide closeness measure. In this setting, a goal is to obtain a reliable characterization of phenotypes, based on very high-dimensional data from a few samples. As opposed to a distance-based approach, the proposed measure defines closeness based on gene signal profiles (functionals) rather than on isolated (numerical) differences in each gene between samples. By using such a measure to characterize intensity differences, we effectively separate biological from artifactual variation in expression, due to tissue effects or signal calibration. Based on this measure, we successively examine the significance of the following: a set of similarly behaved genes relative to all arrayed genes, a set of candidate genes relative to similarly behaved genes, individual candidate genes relative to non-candidates, and the direction, as over- or under-expressed, of candidate genes. In each setting, sample pairs are the units of analysis, with U-statistics the theoretical framework. We illustrate the method on a microarray experiment, where the goal is to select sets of genes that characterize a type of skin cancer and its histological subtypes.     

Bayesian methods for setting sample sizes and choosing allocation ratios in phase II clinical trials with time‐to‐event endpoints

Conventional phase II trials using binary endpoints as early indicators of a time‐to‐event outcome are not always feasible. Uveal melanoma has no reliable intermediate marker of efficacy. In pancreatic cancer and viral clearance, the time to the event of interest is short, making an early indicator unnecessary. In the latter application, Weibull models have been used to analyse corresponding time‐to‐event data. Bayesian sample size calculations are presented for single‐arm and randomised phase II trials assuming proportional hazards models for time‐to‐event endpoints. Special consideration is given to the case where survival times follow the Weibull distribution. The proposed methods are demonstrated through an illustrative trial based on uveal melanoma patient data. A procedure for prior specification based on knowledge or predictions of survival patterns is described. This enables investigation into the choice of allocation ratio in the randomised setting to assess whether a control arm is indeed required. The Bayesian framework enables sample sizes consistent with those used in practice to be obtained. When a confirmatory phase III trial will follow if suitable evidence of efficacy is identified, Bayesian approaches are less controversial than for definitive trials. In the randomised setting, a compromise for obtaining feasible sample sizes is a loss in certainty in the specified hypotheses: the Bayesian counterpart of power. However, this approach may still be preferable to running a single‐arm trial where no data is collected on the control treatment. This dilemma is present in most phase II trials, where resources are not sufficient to conduct a definitive trial. Copyright © 2015 John Wiley & Sons, Ltd.     

A priority rating system for public health programs

When resources are limited, decisions must be made regarding which public health activities to undertake. A priority rating system, which incorporates various data sources, can be used to quantify disease problems or risk factors, or both. The model described in this paper ranks public health issues according to size, urgency, severity of the problem, economic loss, impact on others, effectiveness, propriety, economics, acceptability, legality of solutions, and availability of resources. As examples of how one State can use the model, rankings have been applied to the following health issues: acquired immunodeficiency syndrome, coronary heart disease, injuries from motor vehicle accidents, and cigarette smoking as a risk factor. In this exercise, smoking is the issue with the highest overall priority rating. The model is sensitive to the precision of the data used to develop the rankings and works best for health issues that are not undergoing rapid change. Cost-benefit and cost-effectiveness analyses can be incorporated into the model or used independently in the priority-setting process. Ideally, the model is used in a group setting with six to eight decision makers who represent the primary agency as well as external organizations. Using this method, health agencies, program directors, or community groups can identify the most critical issues or problems requiring intervention programs.     

Statistical considerations for the analysis of massively parallel reporter assays data

Noncoding DNA contains gene regulatory elements that alter gene expression, and the function of these elements can be modified by genetic variation. Massively parallel reporter assays (MPRA) enable high-throughput identification and characterization of functional genetic variants, but the statistical methods to identify allelic effects in MPRA data have not been fully developed. In this study, we demonstrate how the baseline allelic imbalance in MPRA libraries can produce biased results, and we propose a novel, nonparametric, adaptive testing method that is robust to this bias. We compare the performance of this method with other commonly used methods, and we demonstrate that our novel adaptive method controls Type I error in a wide range of scenarios while maintaining excellent power. We have implemented these tests along with routines for simulating MPRA data in the Analysis Toolset for MPRA (@MPRA), an R package for the design and analyses of MPRA experiments. It is publicly available at http://github.com/redaq/atMPRA .     

Novel approaches; improved diagnosis and therapy with DNA microarrays. I. Technology and data analysis

With DNA microarrays it will be possible to refine diagnostics and treatment with the use of genome wide information on a patient's sample. Microarrays are currently applied to unravel gene functions, pathogenetic mechanisms, metabolic routes and the effects of drugs on these, to refine diagnostic classifications and prognostic indexes, and to find new targets for therapy. With genotyping it will be possible to generate risk profiles for certain diseases and to assess the likelihood of a given drug-related side effect. Furthermore, it may accelerate research of mutations in genes for which the normal DNA sequences are already known. The basic principle of DNA microarrays is that thousands of different DNA sequences, each specific for a given gene, are immobilised on a solid surface (for example a microscope slide), arranged in a known order, and are hybridised with a solution of labelled DNA or RNA molecules. The DNA or RNA molecules under investigation bind to complementary base pairs on the slide, and permit us to measure the amount of labelled DNA or RNA hybridised to each gene sequence. Sophisticated software is used to analyse the large amount of data generated. The ultimate aim is to group genes with similar expression patterns across all samples, and then group samples accordingly. Conventional biological or biochemical techniques are required to validate the data obtained with microarrays, and to verify whether the observed associations among genes are biologically relevant.     

Extending the MR‐Egger method for multivariable Mendelian randomization to correct for both measured and unmeasured pleiotropy

Methods have been developed for Mendelian randomization that can obtain consistent causal estimates while relaxing the instrumental variable assumptions. These include multivariable Mendelian randomization, in which a genetic variant may be associated with multiple risk factors so long as any association with the outcome is via the measured risk factors (measured pleiotropy), and the MR‐Egger (Mendelian randomization‐Egger) method, in which a genetic variant may be directly associated with the outcome not via the risk factor of interest, so long as the direct effects of the variants on the outcome are uncorrelated with their associations with the risk factor (unmeasured pleiotropy). In this paper, we extend the MR‐Egger method to a multivariable setting to correct for both measured and unmeasured pleiotropy. We show, through theoretical arguments and a simulation study, that the multivariable MR‐Egger method has advantages over its univariable counterpart in terms of plausibility of the assumption needed for consistent causal estimation and power to detect a causal effect when this assumption is satisfied. The methods are compared in an applied analysis to investigate the causal effect of high‐density lipoprotein cholesterol on coronary heart disease risk. The multivariable MR‐Egger method will be useful to analyse high‐dimensional data in situations where the risk factors are highly related and it is difficult to find genetic variants specifically associated with the risk factor of interest (multivariable by design), and as a sensitivity analysis when the genetic variants are known to have pleiotropic effects on measured risk factors.