Cargando…

Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data

The analyses of large volumes of metagenomic data extracted from aggregate populations of microscopic organisms residing on and in the human body are advancing contemporary understandings of the integrated participation of microbes in human health and disease. Next generation sequencing technology f...

Descripción completa

Detalles Bibliográficos
Autores principales: Pendegraft, Amanda H., Guo, Boyi, Yi, Nengjun
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Public Library of Science 2019
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6706006/
https://www.ncbi.nlm.nih.gov/pubmed/31437194
http://dx.doi.org/10.1371/journal.pone.0220961
_version_ 1783445643986993152
author Pendegraft, Amanda H.
Guo, Boyi
Yi, Nengjun
author_facet Pendegraft, Amanda H.
Guo, Boyi
Yi, Nengjun
author_sort Pendegraft, Amanda H.
collection PubMed
description The analyses of large volumes of metagenomic data extracted from aggregate populations of microscopic organisms residing on and in the human body are advancing contemporary understandings of the integrated participation of microbes in human health and disease. Next generation sequencing technology facilitates said analyses in terms of diversity, community composition, and differential abundance by filtering and binning microbial 16S rRNA genes extracted from human tissues into operational taxonomic units. However, current statistical tools restrict study designs to investigations of limited numbers of host characteristics mediated by limited numbers of samples potentially yielding a loss of relevant information. This paper presents a Bayesian hierarchical negative binomial model as an efficient technique capable of compensating for multivariable sets including tens or hundreds of host characteristics as covariates further expanding analyses of human microbiome count data. Simulation studies reveal that the Bayesian hierarchical negative binomial model provides a desirable strategy by often outperforming three competing negative binomial model in terms of type I error while simultaneously maintaining consistent power. An application of the Bayesian hierarchical negative binomial model using subsets of the open data published by the American Gut Project demonstrates an ability to identify operational taxonomic units significantly differentiable among persons diagnosed by a medical professional with either inflammatory bowel disease or irritable bowel syndrome that are consistent with contemporary gastrointestinal literature.
format Online
Article
Text
id pubmed-6706006
institution National Center for Biotechnology Information
language English
publishDate 2019
publisher Public Library of Science
record_format MEDLINE/PubMed
spelling pubmed-67060062019-09-04 Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data Pendegraft, Amanda H. Guo, Boyi Yi, Nengjun PLoS One Research Article The analyses of large volumes of metagenomic data extracted from aggregate populations of microscopic organisms residing on and in the human body are advancing contemporary understandings of the integrated participation of microbes in human health and disease. Next generation sequencing technology facilitates said analyses in terms of diversity, community composition, and differential abundance by filtering and binning microbial 16S rRNA genes extracted from human tissues into operational taxonomic units. However, current statistical tools restrict study designs to investigations of limited numbers of host characteristics mediated by limited numbers of samples potentially yielding a loss of relevant information. This paper presents a Bayesian hierarchical negative binomial model as an efficient technique capable of compensating for multivariable sets including tens or hundreds of host characteristics as covariates further expanding analyses of human microbiome count data. Simulation studies reveal that the Bayesian hierarchical negative binomial model provides a desirable strategy by often outperforming three competing negative binomial model in terms of type I error while simultaneously maintaining consistent power. An application of the Bayesian hierarchical negative binomial model using subsets of the open data published by the American Gut Project demonstrates an ability to identify operational taxonomic units significantly differentiable among persons diagnosed by a medical professional with either inflammatory bowel disease or irritable bowel syndrome that are consistent with contemporary gastrointestinal literature. Public Library of Science 2019-08-22 /pmc/articles/PMC6706006/ /pubmed/31437194 http://dx.doi.org/10.1371/journal.pone.0220961 Text en © 2019 Pendegraft et al http://creativecommons.org/licenses/by/4.0/ 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.
spellingShingle Research Article
Pendegraft, Amanda H.
Guo, Boyi
Yi, Nengjun
Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
title Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
title_full Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
title_fullStr Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
title_full_unstemmed Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
title_short Bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
title_sort bayesian hierarchical negative binomial models for multivariable analyses with applications to human microbiome count data
topic Research Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6706006/
https://www.ncbi.nlm.nih.gov/pubmed/31437194
http://dx.doi.org/10.1371/journal.pone.0220961
work_keys_str_mv AT pendegraftamandah bayesianhierarchicalnegativebinomialmodelsformultivariableanalyseswithapplicationstohumanmicrobiomecountdata
AT guoboyi bayesianhierarchicalnegativebinomialmodelsformultivariableanalyseswithapplicationstohumanmicrobiomecountdata
AT yinengjun bayesianhierarchicalnegativebinomialmodelsformultivariableanalyseswithapplicationstohumanmicrobiomecountdata