Unsupervised Empirical Bayesian Multiple Testing with External Covariates
نویسندگان
چکیده
In an empirical Bayesian setting, we provide a new multiple testing method, useful when an additional covariate is available, that influences the probability of each null hypothesis being true. We measure the posterior significance of each test conditionally on the covariate and the data, leading to greater power. Using covariate-based prior information in an unsupervised fashion, we produce a list of significant hypotheses which differs in length and order from the list obtained by methods not taking covariate-information into account. Covariate-modulated posterior probabilities of each null hypothesis are estimated using a fast approximate algorithm. The new method is applied to expression quantitative trait loci (eQTL) data.
منابع مشابه
Prior parameter transformation for unsupervised speaker adaptation
In a strictly Bayesian approach, prior parameters are assumed known, based on common or subjective knowledge. But a practical solution for maximum a posteriori adaptation methods is to adopt an empirical Bayesian approach, where the prior parameters are estimated directly from training speech data itself. So there is a problem of mismatches between training and testing conditions in the use of ...
متن کاملThe modeling of body's immune system using Bayesian Networks
In this paper, the urinary infection, that is a common symptom of the decline of the immune system, is discussed based on the well-known algorithms in machine learning, such as Bayesian networks in both Markov and tree structures. A large scale sampling has been executed to evaluate the performance of Bayesian network algorithm. A number of 4052 samples wereobtained from the database of the Tak...
متن کاملUnsupervised Kernel Dimension Reduction
We apply the framework of kernel dimension reduction, originally designed for supervised problems, to unsupervised dimensionality reduction. In this framework, kernel-based measures of independence are used to derive low-dimensional representations that maximally capture information in covariates in order to predict responses. We extend this idea and develop similarly motivated measures for uns...
متن کاملBayesian Frequentist Multiple Testing
We introduce a Bayesian approach to multiple testing. The method is an extension of the false discovery rate (FDR) method due to Benjamini and Hochberg (1995). We also examine the empirical Bayes approach to simultaneous inference proposed by Efron, Tibshirani, Storey and Tusher (2001). We show that, in contrast to the single hypothesis case – where Bayes and frequentist tests do not agree even...
متن کاملMultiple Testing and the Variable-Selection Problem
We study the multiplicity-correction effect of standard Bayesian variable-selection priors in linear regression. Specifically, we compare empirical-Bayes (EB) and fully Bayesian (FB) approaches for handling the prior inclusion probability p required by these priors. Several new information-theoretic results, along with extensive computer experiments, lead us to conclude that the empirical-Bayes...
متن کامل