Empirical Bayesian Research Articles

Identification of key genes for guiding chemotherapeutic management in ovarian cancer using translational bioinformatics

The emergence of resistance to chemotherapy drugs in patients with ovarian cancer is still the main cause of low survival rates. The present study aimed to identify key genes that may provide treatment guidance to reduce the incidence of drug resistance in patients with ovarian cancer. Original data of chemotherapy sensitivity and chemoresistance of ovarian cancer were obtained from the Gene Expression Omnibus dataset GSE73935. Differentially expressed genes (DEGs) between sensitive and resistant ovarian cancer cell lines were screened by Empirical Bayes methods. Overlapping DEGs between four chemoresistant groups were identified by Venn map analysis. Protein-protein interaction networks were also constructed, and hub genes were identified. The hub genes were verified by in vitro experiments as well as The Cancer Genome Atlas data. Results from the present study identified eight important genes that may guide treatment decisions regarding chemotherapy regimens for ovarian cancer, including epidermal growth factor-like repeats and discoidin I-like domains 3, NRAS proto-oncogene, hyaluronan and proteoglycan link protein 1, activated protein C receptor, CD53, cyclin-dependent kinase inhibitor 2A, insulin-like growth factor 1 receptor and roundabout guidance receptor 2 genes. Their expressions were found to have an impact on the prognosis of different treatment groups (cisplatin, paclitaxel, cisplatin + paclitaxel, cisplatin + doxorubicin and cisplatin + topotecan). The results indicated that these genes may minimise the occurrence of ovarian cancer drug resistance and may provide effective treatment options for patients with ovarian cancer.

Scalable Hyperparameter Selection for Latent Dirichlet Allocation

Latent Dirichlet allocation (LDA) is a heavily used Bayesian hierarchical model used in machine learning for modeling high-dimensional sparse count data, for example, text documents. As a Bayesian model, it incorporates a prior on a set of latent variables. The prior is indexed by some hyperparameters, which have a big impact on inference regarding the model. The ideal estimate of the hyperparameters is the empirical Bayes estimate which is, by definition, the maximizer of the marginal likelihood of the data with all the latent variables integrated out. This estimate cannot be obtained analytically. In practice, the hyperparameters are chosen either in an ad-hoc manner, or through some variants of the EM algorithm for which the theoretical basis is weak. We propose an MCMC-based fully Bayesian method for obtaining the empirical Bayes estimate of the hyperparameter. We compare our method with other existing approaches both on synthetic and real data. The comparative experiments demonstrate that the LDA model with hyperparameters specified by our method outperforms models with the hyperparameters estimated by other methods. Supplementary materials for this article are available online.

Safety evaluation of drowsy driving advisory system: Alabama case study

Introduction: The present study discusses roles, characteristics, and safety assessment of a drowsy driving advisory (DDA) system, implemented on rural interstates of Alabama. The DDA system is an engineering countermeasure designed to reduce the likelihood of drowsy driving crashes. It consists of a series of roadside signs with warning and advisory messages for drowsy drivers. The DDA system was implemented upstream of rural rest areas based on a comprehensive crash analysis. Method: A post-implementation study was conducted three years after the DDA system implementation to assess its safety effects. An empirical bayes (EB) method along with predictive methods of the Highway Safety Manual was used in the safety assessment. To overcome the underreported issue of drowsy driving crashes in the crash analysis, the present study used a concept called, expanded definition of drowsy driving (EDD) crashes. Result: The analysis found that the DDA system could reduce total and EDD crashes by 64% and 49%, respectively. It is important to note that such huge crash reduction effects are due to a combined effect of both rest areas and the DDA system, not because of a single treatment. The safety effect of a rest area itself, without considering the effect of the DDA system, was also investigated. Results show that total and EDD crashes would increase about 12–45% and 5–33%, respectively if there is no presence of a rest area. Conclusion: Our findings conclude that the DDA system could significantly reduce both total and drowsy driving crashes when it cooperates with a rest area facility. Practical Application: The findings also provide the guidance of using the DDA system on high-speed roads as a safety countermeasure of drowsy driving crashes. Readers can find details of the DDA system used in this study with its layout, dimension, and roadside safety messages.

Systemic approach to improve safety of urban unsignalized intersections: Development and validation of a Safety Index

Methods based on crash data analysis are effective in identifying intersections with a potential for safety improvement. However, it is well recognized that crash data suffer from several shortcomings and that there are clues to safety other than crash occurrence. The systemic approach is an alternative method to address safety issues. In this approach, a transportation agency is able to identify priority locations based on the presence of risk factors rather than actual crashes. To promote wider use of the systemic safety approach, this paper aims at developing and validating a procedure to rank unsignalised urban intersections for safety improvement based on the evaluation of risk factors by road safety inspections.The procedure assesses a Safety Index (SI) that measures the safety performance of unsignalised urban intersections. The SI is formulated by combining two components of risk: the exposure of road users to road hazards (Exposure) and the risk factors, which increase the probability of involvement in crashes (Risk Index). The SI is made of two elements, one related to vehicles and one to pedestrians. Twenty-three detailed safety issues and ten general safety issues are assessed to compute the vehicle Risk Index and the pedestrian Risk Index. Safety issues were selected considering that they are common issues and that effective remedial measures exist and have already proven their effectiveness. Finally, criteria for identifying and ranking safety issues were defined. The SI has two main practical applications. High risk intersections, where safety measures that can reduce crash frequency exist, can be identified and ranked by the SI. Specific safety issues, that give more contribution to unsafety, are pointed out in order to give indication about more appropriate safety measures according to the systemic safety approach.The procedure was validated with a sample of eighty-nine urban intersections located in Orange County (Florida, U.S.). For these intersections, the SI scores, the Empirical Bayes (EB) safety estimates, and the potential for improvement (PFI) were compared. The correlation between the SI scores and the EB estimates was highly significant both for vehicles (R2 = 0.66) and pedestrians (R2 = 0.58) as well as for the total crashes (R2 = 0.68). The results from the Spearman's rank-correlation analysis provided further validation for the SI indicating that ranking from the SI and the EB estimates agree at the 99.9% confidence level for vehicles (ρs = 0.78), pedestrians (ρs = 0.93), and total (ρs = 0.93).

On the nonparametric maximum likelihood estimator for Gaussian location mixture densities with application to Gaussian denoising

We study the nonparametric maximum likelihood estimator (NPMLE) for estimating Gaussian location mixture densities in $d$-dimensions from independent observations. Unlike usual likelihood-based methods for fitting mixtures, NPMLEs are based on convex optimization. We prove finite sample results on the Hellinger accuracy of every NPMLE. Our results imply, in particular, that every NPMLE achieves near parametric risk (up to logarithmic multiplicative factors) when the true density is a discrete Gaussian mixture without any prior information on the number of mixture components. NPMLEs can naturally be used to yield empirical Bayes estimates of the oracle Bayes estimator in the Gaussian denoising problem. We prove bounds for the accuracy of the empirical Bayes estimate as an approximation to the oracle Bayes estimator. Here our results imply that the empirical Bayes estimator performs at nearly the optimal level (up to logarithmic factors) for denoising in clustering situations without any prior knowledge of the number of clusters.

Roper state from overlap fermions

The Roper state is extracted with valence overlap fermions on a $2+1$-flavor domain-wall fermion lattice (spacing $a = 0.114$ fm and $m_{\pi} = 330$ MeV) using both the Sequential Empirical Bayes (SEB) method and the variational method. The results are consistent, provided that a large smearing-size interpolation operator is included in the variational calculation to have better overlap with the lowest radial excitation. Similar calculations carried out for an anisotropic clover lattice with similar parameters find the Roper $\approx 280$ MeV higher than that of the overlap fermion. The fact that the prediction of the Roper state by overlap fermions is consistently lower than those of clover fermions, chirally improved fermions, and twisted-mass fermions over a wide range of pion masses has been dubbed a "Roper puzzle." To understand the origin of this difference, we study the hairpin $Z$-diagram in the isovector scalar meson ($a_0$) correlator in the quenched approximation. Comparing the $a_0$ correlators for clover and overlap fermions, at a pion mass of 290 MeV, we find that the spectral weight of the ghost state with clover fermions is smaller than that of the overlap at $a = 0.12$ fm and $0.09$ fm, whereas the whole $a_0$ correlators of clover and overlap at $a = 0.06$ fm coincide within errors. This suggests that chiral symmetry is restored for clover at $a \le 0.06$ fm and that the Roper should come down at and below this $a$. We conclude that this work supports a resolution of the "Roper puzzle" due to $Z$-graph type chiral dynamics. This entails coupling to higher components in the Fock space (e.g. $N\pi$, $N\pi\pi$ states) to induce the effective flavor-spin interaction between quarks as prescribed in the chiral quark model, resulting in the parity-reversal pattern as observed in the experimental excited states of $N, \Delta$ and $\Lambda$.

Performance of empirical Bayes estimation techniques used in probabilistic risk assessment

This paper analyzes, validates, compares, and contrasts parameter estimation techniques used in probabilistic risk assessment. The analysis and validation are performed using computer-simulated as well as real-world component reliability data collected in nuclear industry. The analysis is focused on parametric and nonparametric empirical Bayes as well as frequentists methods. Different techniques are analyzed in terms of total squared error risk. A recently published nonparametric Bayes method based on the solution of an integral equation is described and its performance is contrasted with parametric and hierarchical Bayes. The results show that the nonparametric Bayes can achieve a smaller squared error risk than other Bayesian methods and maximum likelihood estimators when applied to the data collected at different nuclear power plants. However, the improvement produced by empirical Bayes techniques is not uniform and cannot be guaranteed for each and every plant. The paper shows strengths and limitations of different estimations techniques and concludes with practical recommendations.

Development of crash modification factors for intersections in Toowoomba city

ABSTRACT This research outlines the analytical approaches that may be employed to recommend multiple treatments for application to improve safety level at hazardous intersections in regional areas. Data from 106 intersections in the regional city of Toowoomba, Queensland, Australia were used in modelling and subsequent analysis. Specifically, crash prediction models were developed to estimate crash modification factors that could be used to propose remedial measures. A total of sixteen geometric and operational conditions were used as explanatory variables including the number of legs at an intersection, the availability of entering and exiting via through lanes, the presence of left-turning lanes, right-turning lanes, slip lanes, median islands, traffic control, and speed limits on the feed roads. The Empirical Bayes approach was employed to identify five hazardous intersections in Toowoomba for further investigation. The four most suitable techniques for estimating combined crash modification factors were reviewed and then used to propose effective road safety measures for the five selected hazardous intersections. Finally, this research suggests four future areas worthy of further investigation. Highlights Estimated the number of road crashes at intersections using crash prediction models. Applied the Empirical Bayes method to identify the most hazardous intersections. Combined crash modifications factors were calculated using various techniques. Identified the effective countermeasures to reduce the road crashes at hazardous intersections.

A generalized framework for classical test theory

This paper develops a generalized framework which allows for the use of parametric classical test theory inference with non-normal models. Using the theory of natural exponential families and Bayesian theory of their conjugate priors, theoretical properties of test scores under the framework are derived, including a formula for parallel-test reliability in terms of the test length and a parameter of the underlying population distribution of abilities. This framework is shown to satisfy the general properties of classical test theory several common classical test theory results are shown to reduce to parallel-test reliability in this framework. An empirical Bayes method for estimating reliability, both with point estimates and with intervals, is described using maximum likelihood. This method is applied to an example set of data and compared to classical test theory estimators of reliability, and a simulation study is performed to show the coverage of the interval estimates of reliability derived from the framework.

Before-and-After Empirical Bayes Evaluation of Citywide Installation of Driver Feedback Signs

Speeding is a leading factor that contributes to approximately one-third of all fatal collisions. Over the past decades, various passive/active countermeasures have been adopted to improve drivers’ compliance to posted speed limits to improve traffic safety. The driver feedback sign (DFS) is considered a low-cost innovative intervention that is being widely used, in growing numbers, in urban cities to provide positive guidance for motorists. Despite their documented effectiveness in reducing speeds, limited literature exists on their impact on reducing collisions. This study addresses this gap by designing a before-and-after study using the empirical Bayes method for a large sample of urban road segments. Safety performance functions and yearly calibration factors are developed to quantify the sole effectiveness of DFS using large-scale spatial data and a set of reference road segments within the city of Edmonton, Alberta, Canada. Likewise, the study followed a detailed economic analysis based on three collision-costing criteria to investigate if DFS was indeed a cost-effective intervention. The results showed significant collision reductions that ranged from 32.5% to 44.9%, with the highest reductions observed for severe speed-related collisions. The results further attested that the benefit–cost ratios, combining severe and property-damage-only collisions, ranged from 8.2 to 20.2 indicating that DFS can be an extremely economical countermeasure. The findings from this study can provide transportation agencies in need of implementing cost-efficient countermeasures with a tool they need to design a long-term strategic deployment plan to ensure the safety of traveling public.

Infectious mononucleosis, immune genotypes, and non-Hodgkin lymphoma (NHL): an InterLymph Consortium study.

We explored the interaction between non-Hodgkin lymphoma (NHL), infectious mononucleosis (IM) history, and immune-related genotypes in a pooled case-control analysis. A total of 7,926 NHL patients and 10,018 controls from 12 case-control studies were included. Studies were conducted during various time periods between 1988 and 2008, and participants were 17-96years of age at the time of ascertainment/recruitment. Self-reported IM history and immune response genotypes were provided by the InterLymph Data Coordinating Center at Mayo Clinic. Odds ratios (OR) were estimated using multivariate logistic regression, and interactions were estimated using the empirical Bayes method. PACT was used to account for multiple comparisons. There was evidence of an interaction effect between IM history and two variants on T-cell lymphoma (TCL) risk: rs1143627 in interleukin-1B (IL1B) (pinteraction = 0.04, ORinteraction=0.09, 95% confidence interval [CI] 0.01, 0.87) and rs1800797 in interleukin-6 (IL6) (pinteraction=0.03, ORinteraction = 0.08, 95% CI 0.01, 0.80). Neither interaction effect withstood adjustment for multiple comparisons. There were no statistically significant interactions between immune response genotypes and IM on other NHL subtypes. Genetic risk variants in IL1B and IL6 may affect the association between IM and TCL, possibly by influencing T-cell activation, growth, and differentiation in the presence of IM, thereby decreasing risk of immune cell proliferation.

Evaluation of Section Speed Enforcement System Using Empirical Bayes Approach and Turning Point Analysis

Speeding is a major risk factor for traffic-related injuries. As a countermeasure against speeding, automated speed enforcement systems (ASES) have been deployed in many countries. However, drivers’ awareness of enforcement locations allows themselves to adjust vehicle speeds in the vicinity of the enforcement locations. This enforcement avoidance behavior leads to a criticism of the effectiveness of ASES, in which the system promotes abrupt changes in vehicle speed near enforcement locations, increasing crash risk as a side effect. To address this issue, the section speed enforcement system (SSES), which enforces overspeeding vehicles by their average travel speed over a section, has been devised. In this study, we evaluate traffic speed and safety data that were collected from sections with SSES on Korean expressways. The speed analysis showed that the vehicles reduced their speeds inside the enforcement section, and this reduction in speed variations across vehicles was also noticeable, signifying that the risk of traffic crash should be lower. In view of this, we have performed before and after comparative analysis using the empirical Bayes method with the comparison group. The outcomes estimate 43% reduction in crash occurrence after installation of SSES. Furthermore, turning point analysis confirmed that the reduction in crash occurrence ensued immediately after installation of SSES.

Spatial clustering of suicide mortality and associated community characteristics in Kanagawa prefecture, Japan, 2011\u20132017

BackgroundSuicide mortality is high in Japan and early interventional strategies to solve that problem are needed. An accurate evaluation of the regional status of current suicide mortality would be useful for community interventions. A few studies in Kanagawa prefecture, located next to Tokyo and with the second largest population in Japan, have identified spatial clusters of suicide mortality at regional levels. This study examined spatial clustering and clustering over time of such events using spatial data from regional statistics on suicide deaths.MethodsData were obtained from regional statistics (58 regions in Kanagawa prefecture) of the National Vital Statistics of Japan from 2011 to 2017. The standardized mortality ratio (SMR) and Empirical Bayes estimator for the SMR (EBSMR) were used as measures. Spatial clusters were examined by Kulldorff’s circular spatial scan statistic, Tango-Takahashi’s flexible spatial scan statistic and Tango’s test. Linear regression and conditional autoregressive (CAR) models were used not only to adjust for covariates but also to estimate regional effects. The analyses were conducted for each year, inclusive.ResultsAmong male suicide deaths, being unemployed (50%) was most frequently related to suicide while among female health problem (50%) were frequent. Spatial clusters with significance detected by FlexScan, SatScan and Tango’s test were few and varied somewhat according to the method used. Spatial clusters were detected in some regions including Kawasaki ward after adjustment by covariates. By the linear regression models, selected variables with significance were different between the sexes. For males, unemployment, family size, and proportion of higher education were detected for several of the years studied while for females, family size and divorce rate were detected over this period. These variables were also observed by the CAR model with 5 covariates. Regional effects were much clearer by considering the spatial parameter for both males and females and especially, Kawasaki ward was detected as a high risk region in many years.ConclusionThe present results detected some spatial clustering of suicide deaths within certain regions. Factors related to suicide deaths were also indicated. These results would provide important information in policy making for suicide prevention.

Lessons learned from the large-scale application of Driver Feedback Signs in an urban city

The City of Edmonton has invested in the installation of Driver Feedback Signs (DFSs) throughout the city starting from 2011. DFSs are dynamic speed display signs aimed at providing positive guidance to drivers with the goal of improving compliance to posted speed limits. Given the city’s extensive history with DFS installation, the goal of this study is to evaluate the safety performance of DFSs and to identify factors that can help in determining the future DFS sites selection. A before-and-after evaluation with Empirical Bayes (EB) adjustment was used to account for regression-to-mean bias and other confounding factors. Local safety performance functions and yearly calibration factors were developed using data from a set of reference urban roads. The EB method analysis was utilized to investigate the effect of DFS on different road and intervention types. Results showed significant collision reductions in all scenarios ranging from 31.0% to 41.6%. DFSs were more effective in reducing collisions for arterials compared to collectors. Also, the combined use of DFS and mobile photo enforcement had a slightly higher effect on safety. Initial collision frequencies, traffic volumes, road lengths and the presence of shoulders were found to impact the reduction in collisions for most collision types.

Application of the Poisson-Tweedie distribution in analyzing crash frequency data

This paper describes a study that applies the Poisson-Tweedie distribution in developing crash frequency models. The Poisson-Tweedie distribution offers a unified framework to model overdispersed, underdispersed, zero-inflated, spatial, and longitudinal count data, as well as multiple response variables of similar or mixed types. The form of its variance function is simple, and can be specified as the mean added to the product of dispersion and mean raised to the power P. The flexibility of the Poisson-Tweedie distribution lies in the domain of P, which includes positive real number values. Special cases of the Poisson-Tweedie distribution models include the linear form of the negative binomial (NB1) model with P equal to 1.0, the geometric Poisson (GeoP) model with P equal to 1.5, the quadratic form of the negative binomial (NB2) model with P equal to 2.0, and the Poisson Inverse Gaussian (PIG) model with P equal to 3.0. A series of models were developed in this study using the Poisson-Tweedie distribution without any restrictions on the value of the power parameter as well as with specific values of the power parameter representing NB1, GeoP, NB2, and PIG models. The effects of fixed and varying dispersion parameters (i.e., dispersion as a function of covariates) on the variance and expected crash frequency estimates were also examined. Three years (2012–2014) of crash data from urban three-leg stop-controlled intersections and urban four-leg signalized intersections in the state of Florida were used to develop the models. The Poisson-Tweedie models or the GeoP models were found to perform better when the dispersion parameter was constant or fixed. With the varying dispersion parameter, the NB2 and PIG models were found to perform better, with both performing equally well. Also, the fixed dispersion parameter values were found to be smaller in the models with a higher value of the power parameter. The variation across the models in their estimates of weight factor, expected crash frequency, and potential for safety improvement of hazardous sites based on the empirical Bayes method was also discussed.

Application of Reticulocyte-Based Estimation of Red Blood Cell Lifespan in Anemia Management of End-Stage Renal Disease Patients.

Shortened red blood cell (RBC) lifespan is one of the major factors contributing to anemia in end-stage renal disease (ESRD) patients and should be taken into account in anemia management protocols. In this study, we aimed to estimate RBC lifespan and the source of between-subject variability in ESRD patients. The resulting individual parameters (empirical Bayes estimates) were used to predict hemoglobin concentrations 2weeks in advance. The reticulocyte-based estimation of RBC lifespan (REBEL) and the population modeling of RBC count data were used. A total of 120 blood samples collected biweekly over 10weeks in 24 patients receiving maintenance doses of recombinant human erythropoietin (rHuEPO) subcutaneously were included in this analysis. Typical RBC lifespan was estimated to be 63.3days. RBC lifespan was found to increase with erythroferrone, a recently identified hormone participating in iron metabolism. Approximately, a 10% increase in plasma erythroferrone was associated with a 5% increase in RBC lifespan. In addition, RBC lifespan was 18.7% shorter in females compared with males. Out of 24 subjects, 16 had hemoglobin concentrations predicted within 95% prediction intervals. The median absolute prediction error was 15.9% (interquartile range, 9.5 to 24.7%). We demonstrated that REBEL coupled with the population modeling technique can be used effectively to estimate RBC lifespan. Then, individual parameters can be used to predict future hemoglobin concentrations in ESRD patients.

Comparative study for safety performance measures of signalized intersection sites

The decision on which safety performance measure to use for diagnosis of sites with probable to profit from safety improvements is a major factor in making the right decision for the guidance of resources. This study concentrated on introducing a quantitative comparison of sites according to four crash-based methods suggested in the Highway Safety Manual (HSM). Archived crash data for three years are obtained to conduct a comparison course. As a case study, the analysis and measure of safety for 9 four-leg signalized intersections in Baghdad city have been done independently using Empirical Bayes method (EB-method), observed Crash Frequency (CF), Crash Rate (CR) and Empirical Bayes Adjustment (EB-adj.) as safety performance measures in HSM. In this paper the EB-method is used as a benchmark for comparison. The safety measures are evaluated through rank correlation analysis while hazard location identification results are compared through the use of values rank-based mean absolute error. Quantitative evaluation tests showed that each of EB-adj and observed CF correlated well with EB-method while CR method exhibited poor performance in comparison with EB-method and was the worst in hazard location identification. This result is quite confusing since many agencies still depend on the CR method in traffic safety analyses.

Optimal prediction in the linearly transformed spiked model

We consider the linearly transformed spiked model, where the observations $Y_{i}$ are noisy linear transforms of unobserved signals of interest $X_{i}$: \begin{equation*}Y_{i}=A_{i}X_{i}+\varepsilon_{i},\end{equation*} for $i=1,\ldots ,n$. The transform matrices $A_{i}$ are also observed. We model the unobserved signals (or regression coefficients) $X_{i}$ as vectors lying on an unknown low-dimensional space. Given only $Y_{i}$ and $A_{i}$ how should we predict or recover their values? The naive approach of performing regression for each observation separately is inaccurate due to the large noise level. Instead, we develop optimal methods for predicting $X_{i}$ by “borrowing strength” across the different samples. Our linear empirical Bayes methods scale to large datasets and rely on weak moment assumptions. We show that this model has wide-ranging applications in signal processing, deconvolution, cryo-electron microscopy, and missing data with noise. For missing data, we show in simulations that our methods are more robust to noise and to unequal sampling than well-known matrix completion methods.

Robust empirical Bayes small area estimation with density power divergence

Summary A two-stage normal hierarchical model called the Fay–Herriot model and the empirical Bayes estimator are widely used to obtain indirect and model-based estimates of means in small areas. However, the performance of the empirical Bayes estimator can be poor when the assumed normal distribution is misspecified. This article presents a simple modification that makes use of density power divergence and proposes a new robust empirical Bayes small area estimator. The mean squared error and estimated mean squared error of the proposed estimator are derived based on the asymptotic properties of the robust estimator of the model parameters. We investigate the numerical performance of the proposed method through simulations and an application to survey data.

Small area estimation with spatially varying natural exponential families

Two-stage hierarchical models have been widely used in small area estimation to produce indirect estimates of areal means. When the areas are treated exchangeably and the model parameters are assumed to be the same over all areas, we might lose the efficiency in the presence of spatial heterogeneity. To overcome this problem, we consider a two-stage area-level model based on natural exponential family with spatially varying model parameters. We employ geographically weighted regression approach to estimating the varying parameters and suggest a new empirical Bayes estimator of the areal mean. We also discuss some related problems, including the mean squared error estimation, benchmarked estimation and estimation in non-sampled areas. The performance of the proposed method is evaluated through simulations and applications to two data sets.