Applied immuno-epidemiological research: an approach for integrating existing knowledge into the statistical analysis of multiple immune markers

BMC Immunology

Methodology Article

Marcus D. Saemann¹  Philip J. Cooper²  Thomas Weichhart³  Camila A. Figueiredo⁴  Neuza Alcântara-Neves⁴  Mauricio L. Barreto⁵  Leila D. Amorim⁶  Bernd Genser⁷  Laura C. Rodrigues⁸  Joachim E. Fischer⁹ 

[1] Clinical Division of Nephrology, Internal Medicine III, Medical University of Vienna, Vienna, Austria;Institute of Infection and Immunity, St George’s University of London, London, UK;Facultad de Ciencias Medicas, de la Salud y la Vida, Universidad Internacional del Ecuador, Quito, Ecuador;Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria;Instituto de Ciências da Saúde, Federal University of Bahia, Salvador, Brazil;Instituto de Saúde Coletiva, Federal University of Bahia, Rua Basílio da Gama, s/n - Canela, 40110-040, Salvador, BA, Brazil;Centro de Pesquisa Gonçalo Muniz, Fundação Oswaldo Cruz (FIOCRUZ), Salvador, Brazil;Instituto de Saúde Coletiva, Federal University of Bahia, Rua Basílio da Gama, s/n - Canela, 40110-040, Salvador, BA, Brazil;Instituto de Matemática, Federal University of Bahia, Salvador, Brazil;Instituto de Saúde Coletiva, Federal University of Bahia, Rua Basílio da Gama, s/n - Canela, 40110-040, Salvador, BA, Brazil;Mannheim Institute of Public Health, Social and Preventive Medicine, University of Heidelberg, Heidelberg, Germany;London School of Hygiene & Tropical Medicine, London, UK;Mannheim Institute of Public Health, Social and Preventive Medicine, University of Heidelberg, Heidelberg, Germany;

关键词: Immuno-epidemiology;    Correlated immune markers;    Cytokines;    Statistical analysis;    Conceptual frameworks;   

DOI  :  10.1186/s12865-016-0149-9

 received in 2015-08-14, accepted in 2016-05-08,  发布年份 2016

来源: Springer

【 摘 要 】

BackgroundImmunologists often measure several correlated immunological markers, such as concentrations of different cytokines produced by different immune cells and/or measured under different conditions, to draw insights from complex immunological mechanisms. Although there have been recent methodological efforts to improve the statistical analysis of immunological data, a framework is still needed for the simultaneous analysis of multiple, often correlated, immune markers. This framework would allow the immunologists’ hypotheses about the underlying biological mechanisms to be integrated.ResultsWe present an analytical approach for statistical analysis of correlated immune markers, such as those commonly collected in modern immuno-epidemiological studies. We demonstrate i) how to deal with interdependencies among multiple measurements of the same immune marker, ii) how to analyse association patterns among different markers, iii) how to aggregate different measures and/or markers to immunological summary scores, iv) how to model the inter-relationships among these scores, and v) how to use these scores in epidemiological association analyses. We illustrate the application of our approach to multiple cytokine measurements from 818 children enrolled in a large immuno-epidemiological study (SCAALA Salvador), which aimed to quantify the major immunological mechanisms underlying atopic diseases or asthma. We demonstrate how to aggregate systematically the information captured in multiple cytokine measurements to immunological summary scores aimed at reflecting the presumed underlying immunological mechanisms (Th1/Th2 balance and immune regulatory network). We show how these aggregated immune scores can be used as predictors in regression models with outcomes of immunological studies (e.g. specific IgE) and compare the results to those obtained by a traditional multivariate regression approach.ConclusionThe proposed analytical approach may be especially useful to quantify complex immune responses in immuno-epidemiological studies, where investigators examine the relationship among epidemiological patterns, immune response, and disease outcomes.

【 授权许可】

CC BY   
© Genser et al. 2016