IRUS Total

Accounting for measurement error to assess the effect of air pollution on omics signals

File Description SizeFormat 
journal.pone.0226102.pdfPublished version2.38 MBAdobe PDFView/Open
Title: Accounting for measurement error to assess the effect of air pollution on omics signals
Authors: Ponzi, E
Vineis, P
Chung, K
Blangiardo, M
Item Type: Journal Article
Abstract: Studies on the effects of air pollution and more generally environmental exposures onhealth require measurements of pollutants, which are affected by measurement error.This is a cause of bias in the estimation of parameters relevant to the study and canlead to inaccurate conclusions when evaluating associations among pollutants, diseaserisk and biomarkers. Although the presence of measurement error in such studies hasbeen recognized as a potential problem, it is rarely considered in applications andpractical solutions are still lacking. In this work, we formulate Bayesian measurementerror models and apply them to study the link between air pollution and omic signals.The data we use stem from the “Oxford Street II Study”, a randomized crossover trialin which 60 volunteers walked for two hours in a traffic-free area (Hyde Park) and in abusy shopping street (Oxford Street) of London. Metabolomic measurements were madein each individual as well as air pollution measurements, in order to investigate theassociation between short-term exposure to traffic related air pollution and perturbationof metabolic pathways. We implemented error-corrected models in a classical frameworkand used the flexibility of Bayesian hierarchical models to account for dependenciesamong omic signals, as well as among different pollutants. Models were implementedusing traditional Markov Chain Monte Carlo (MCMC) simulative methods as well asintegrated Laplace approximation. The inclusion of a classical measurement error termresulted in variable estimates of the association between omic signals and traffic relatedair pollution measurements, where the direction of the bias was not predictable a priori.The models were successful in including and accounting for different correlationstructures, both among omic signals and among different pollutant exposures. Ingeneral, more associations were identified when the correlation among omics and amongpollutants were modeled, and their number increased when a measurement error termwas additionally included in the multivariate models (particularly for the associationsbetween metabolomics andNO2).
Issue Date: 2-Jan-2020
Date of Acceptance: 19-Nov-2019
URI: http://hdl.handle.net/10044/1/75701
DOI: 10.1371/journal.pone.0226102
ISSN: 1932-6203
Publisher: Public Library of Science (PLoS)
Start Page: 1
End Page: 16
Journal / Book Title: PLoS One
Volume: 15
Issue: 1
Copyright Statement: © 2020 Ponzi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Sponsor/Funder: Commission of the European Communities
Funder's Grant Number: 308610
Keywords: Air Pollution
Bayes Theorem
Markov Chains
Models, Statistical
Monte Carlo Method
Time Factors
Models, Statistical
Monte Carlo Method
Bayes Theorem
Markov Chains
Air Pollution
Time Factors
General Science & Technology
Publication Status: Published
Online Publication Date: 2020-01-02
Appears in Collections:National Heart and Lung Institute
Grantham Institute for Climate Change