Annals of Applied Statistics最新文献

Wildland fire smoke contains hazardous levels of fine particulate matter (PM_2.5), a pollutant shown to adversely effect health. Estimating fire attributable PM_2.5 concentrations is key to quantifying the impact on air quality and subsequent health burden. This is a challenging problem since only total PM_2.5 is measured at monitoring stations and both fire-attributable PM_2.5 and PM_2.5 from all other sources are correlated in space and time. We propose a framework for estimating fire-contributed PM_2.5 and PM_2.5 from all other sources using a novel causal inference framework and bias-adjusted chemical model representations of PM_2.5 under counterfactual scenarios. The chemical model representation of PM_2.5 for this analysis is simulated using Community Multiscale Air Quality Modeling System (CMAQ), run with and without fire emissions across the contiguous U.S. for the 2008-2012 wildfire seasons. The CMAQ output is calibrated with observations from monitoring sites for the same spatial domain and time period. We use a Bayesian model that accounts for spatial variation to estimate the effect of wildland fires on PM_2.5 and state assumptions under which the estimate has a valid causal interpretation. Our results include estimates of the contributions of wildfire smoke to PM_2.5 for the contiguous U.S. Additionally, we compute the health burden associated with the PM_2.5 attributable to wildfire smoke.

We propose the "study strap ensemble", which combines advantages of two common approaches to fitting prediction models when multiple training datasets ("studies") are available: pooling studies and fitting one model versus averaging predictions from multiple models each fit to individual studies. The study strap ensemble fits models to bootstrapped datasets, or "pseudo-studies." These are generated by resampling from multiple studies with a hierarchical resampling scheme that generalizes the randomized cluster bootstrap. The study strap is controlled by a tuning parameter that determines the proportion of observations to draw from each study. When the parameter is set to its lowest value, each pseudo-study is resampled from only a single study. When it is high, the study strap ignores the multi-study structure and generates pseudo-studies by merging the datasets and drawing observations like a standard bootstrap. We empirically show the optimal tuning value often lies in between, and prove that special cases of the study strap draw the merged dataset and the set of original studies as pseudo-studies. We extend the study strap approach with an ensemble weighting scheme that utilizes information in the distribution of the covariates of the test dataset. Our work is motivated by neuroscience experiments using real-time neurochemical sensing during awake behavior in humans. Current techniques to perform this kind of research require measurements from an electrode placed in the brain during awake neurosurgery and rely on prediction models to estimate neurotransmitter concentrations from the electrical measurements recorded by the electrode. These models are trained by combining multiple datasets that are collected in vitro under heterogeneous conditions in order to promote accuracy of the models when applied to data collected in the brain. A prevailing challenge is deciding how to combine studies or ensemble models trained on different studies to enhance model generalizability. Our methods produce marked improvements in simulations and in this application. All methods are available in the studyStrap CRAN package.

Identifying differences in networks has become a canonical problem in many biological applications. Existing methods try to accomplish this goal by either directly comparing the estimated structures of two networks, or testing the null hypothesis that the covariance or inverse covariance matrices in two populations are identical. However, estimation approaches do not provide measures of uncertainty, e.g., p-values, whereas existing testing approaches could lead to misleading results, as we illustrate in this paper. To address these shortcomings, we propose a qualitative hypothesis testing framework, which tests whether the connectivity structures in the two networks are the same. our framework is especially appropriate if the goal is to identify nodes or edges that are differentially connected. No existing approach could test such hypotheses and provide corresponding measures of uncertainty. Theoretically, we show that under appropriate conditions, our proposal correctly controls the type-I error rate in testing the qualitative hypothesis. Empirically, we demonstrate the performance of our proposal using simulation studies and applications in cancer genomics.

Understanding sub-cellular protein localisation is an essential component in the analysis of context specific protein function. Recent advances in quantitative mass-spectrometry (MS) have led to high resolution mapping of thousands of proteins to sub-cellular locations within the cell. Novel modelling considerations to capture the complex nature of these data are thus necessary. We approach analysis of spatial proteomics data in a non-parametric Bayesian framework, using K-component mixtures of Gaussian process regression models. The Gaussian process regression model accounts for correlation structure within a sub-cellular niche, with each mixture component capturing the distinct correlation structure observed within each niche. The availability of marker proteins (i.e. proteins with a priori known labelled locations) motivates a semi-supervised learning approach to inform the Gaussian process hyperparameters. We moreover provide an efficient Hamiltonian-within-Gibbs sampler for our model. Furthermore, we reduce the computational burden associated with inversion of covariance matrices by exploiting the structure in the covariance matrix. A tensor decomposition of our covariance matrices allows extended Trench and Durbin algorithms to be applied to reduce the computational complexity of inversion and hence accelerate computation. We provide detailed case-studies on Drosophila embryos and mouse pluripotent embryonic stem cells to illustrate the benefit of semi-supervised functional Bayesian modelling of the data.

Late-stage clinical trials have been conducted primarily to establish the efficacy of a new treatment in an intended population. A corollary of population heterogeneity in clinical trials is that a treatment might be effective for one or more subgroups, rather than for the whole population of interest. As an example, the phase III clinical trial of panitumumab in metastatic colorectal cancer patients failed to demonstrate its efficacy in the overall population, but a subgroup associated with tumor KRAS status was found to be promising (Peeters et al. (Am. J. Clin. Oncol. 28 (2010) 4706-4713)). As we search for such subgroups via data partitioning based on a large number of biomarkers, we need to guard against inflated type I error rates due to multiple testing. Commonly-used multiplicity adjustments tend to lose power for the detection of subgroup treatment effects. We develop an effective omnibus test to detect the existence of, at least, one subgroup treatment effect, allowing a large number of possible subgroups to be considered and possibly censored outcomes. Applied to the panitumumab trial data, the proposed test would confirm a significant subgroup treatment effect. Empirical studies also show that the proposed test is applicable to a variety of outcome variables and maintains robust statistical power.

Understanding the role of time-varying pollution mixtures on human health is critical as people are simultaneously exposed to multiple pollutants during their lives. For vulnerable subpopulations who have well-defined exposure periods (e.g., pregnant women), questions regarding critical windows of exposure to these mixtures are important for mitigating harm. We extend critical window variable selection (CWVS) to the multipollutant setting by introducing CWVS for mixtures (CWVSmix), a hierarchical Bayesian method that combines smoothed variable selection and temporally correlated weight parameters to: (i) identify critical windows of exposure to mixtures of time-varying pollutants, (ii) estimate the time-varying relative importance of each individual pollutant and their first order interactions within the mixture, and (iii) quantify the impact of the mixtures on health. Through simulation we show that CWVSmix offers the best balance of performance in each of these categories in comparison to competing methods. Using these approaches, we investigate the impact of exposure to multiple ambient air pollutants on the risk of stillbirth in New Jersey, 2005-2014. We find consistent elevated risk in gestational weeks 2, 16-17, and 20 for non-Hispanic Black mothers, with pollution mixtures dominated by ammonium (weeks 2, 17, 20), nitrate (weeks 2, 17), nitrogen oxides (weeks 2, 16), PM_2.5 (week 2), and sulfate (week 20). The method is available in the R package CWVSmix.

To combat the HIV/AIDS pandemic effectively, targeted interventions among certain key populations play a critical role. Examples of such key populations include sex workers, people who inject drugs, and men who have sex with men. While having accurate estimates for the size of these key populations is important, any attempt to directly contact or count members of these populations is difficult. As a result, indirect methods are used to produce size estimates. Multiple approaches for estimating the size of such populations have been suggested but often give conflicting results. It is, therefore, necessary to have a principled way to combine and reconcile these estimates. To this end, we present a Bayesian hierarchical model for estimating the size of key populations that combines multiple estimates from different sources of information. The proposed model makes use of multiple years of data and explicitly models the systematic error in the data sources used. We use the model to estimate the size of people who inject drugs in Ukraine. We evaluate the appropriateness of the model and compare the contribution of each data source to the final estimates.

In high-dimensional regression problems, often a relatively small subset of the features are relevant for predicting the outcome, and methods that impose sparsity on the solution are popular. When multiple correlated outcomes are available (multitask), reduced rank regression is an effective way to borrow strength and capture latent structures that underlie the data. Our proposal is motivated by the UK Biobank population-based cohort study, where we are faced with large-scale, ultrahigh-dimensional features, and have access to a large number of outcomes (phenotypes)-lifestyle measures, biomarkers, and disease outcomes. We are hence led to fit sparse reduced-rank regression models, using computational strategies that allow us to scale to problems of this size. We use a scheme that alternates between solving the sparse regression problem and solving the reduced rank decomposition. For the sparse regression component we propose a scalable iterative algorithm based on adaptive screening that leverages the sparsity assumption and enables us to focus on solving much smaller subproblems. The full solution is reconstructed and tested via an optimality condition to make sure it is a valid solution for the original problem. We further extend the method to cope with practical issues, such as the inclusion of confounding variables and imputation of missing values among the phenotypes. Experiments on both synthetic data and the UK Biobank data demonstrate the effectiveness of the method and the algorithm. We present multiSnpnet package, available at http://github.com/junyangq/multiSnpnet that works on top of PLINK2 files, which we anticipate to be a valuable tool for generating polygenic risk scores from human genetic studies.

Functional magnetic resonance imaging (fMRI) has provided invaluable insight into our understanding of human behavior. However, large inter-individual differences in both brain anatomy and functional localization after anatomical alignment remain a major limitation in conducting group analyses and performing population level inference. This paper addresses this problem by developing and validating a new computational technique for reducing misalignment across individuals in functional brain systems by spatially transforming each subjects functional data to a common reference map. Our proposed Bayesian functional registration approach allows us to assess differences in brain function across subjects and individual differences in activation topology. It combines intensity-based and feature-based information into an integrated framework, and allows inference to be performed on the transformation via the posterior samples. We evaluate the method in a simulation study and apply it to data from a study of thermal pain. We find that the proposed approach provides increased sensitivity for group-level inference.

This article addresses the evaluation of post-randomization immune response biomarkers as principal surrogate endpoints of a vaccine's protective effect, based on data from randomized vaccine trials. An important metric for quantifying a biomarker's principal surrogacy in vaccine research is the vaccine efficacy curve, which shows a vaccine's efficacy as a function of potential biomarker values if receiving vaccine, among an 'early-always-at-risk' principal stratum of trial participants who remain disease-free at the time of biomarker measurement whether having received vaccine or placebo. Earlier work in principal surrogate evaluation relied on an 'equal-early-clinical-risk' assumption for identifiability of the vaccine curve, based on observed disease status at the time of biomarker measurement. This assumption is violated in the common setting that the vaccine has an early effect on the clinical endpoint before the biomarker is measured. In particular, a vaccine's early protective effect observed in two phase III dengue vaccine trials (CYD14/CYD15) has motivated our current research development. We relax the 'equal-early-clinical-risk' assumption and propose a new sensitivity analysis framework for principal surrogate evaluation allowing for early vaccine efficacy. Under this framework, we develop inference procedures for vaccine efficacy curve estimators based on the estimated maximum likelihood approach. We then use the proposed methodology to assess the surrogacy of post-randomization neutralization titer in the motivating dengue application.