Back to Search
Start Over
A systematic machine learning and data type comparison yields metagenomic predictors of infant age, sex, breastfeeding, antibiotic usage, country of origin, and delivery type
- Source :
- PLoS Computational Biology, PLoS Computational Biology, Vol 16, Iss 5, p e1007895 (2020)
- Publication Year :
- 2020
- Publisher :
- Public Library of Science, 2020.
-
Abstract
- The microbiome is a new frontier for building predictors of human phenotypes. However, machine learning in the microbiome is fraught with issues of reproducibility, driven in large part by the wide range of analytic models and metagenomic data types available. We aimed to build robust metagenomic predictors of host phenotype by comparing prediction performances and biological interpretation across 8 machine learning methods and 4 different types of metagenomic data. Using 1,570 samples from 300 infants, we fit 7,865 models for 6 host phenotypes. We demonstrate the dependence of accuracy on algorithm choice and feature definition in microbiome data and propose a framework for building microbiome-derived indicators of host phenotype. We additionally identify biological features predictive of age, sex, breastfeeding status, historical antibiotic usage, country of origin, and delivery type. Our complete results can be viewed at http://apps.chiragjpgroup.org/ubiome_predictions/.<br />Author summary The human microbiome is hypothesized to influence human phenotype. However, many published host-microbe associations may not be reproducible. A number of reasons could be behind irreproducible results, including a wide array of methods for measuring the microbiome through genetic sequence, annotation pipelines, and analytical models/prediction approaches. Therefore, there is a need to compare different modeling strategies and microbiome data types (i.e. species abundance versus metabolic pathway abundance) to determine how to build robust and reproducible host-microbiome predictions. In this work, we executed a broad comparison of different predictive methods as a function of microbiome data types to effectively predict host characteristics. Our pipeline was able uncover robust microbial associations with phenotype. We additionally recommended considerations for reproducible microbiome-host association pipeline development. We claim our work is a necessary stepping stone in increasing the utility of emerging cohort data and enabling the next generation of efficient microbiome association studies in human health.
- Subjects :
- 0301 basic medicine
Male
Maternal Health
Breastfeeding
computer.software_genre
Pediatrics
Machine Learning
0302 clinical medicine
Mathematical and Statistical Techniques
Antibiotics
Feature (machine learning)
Medicine and Health Sciences
Biology (General)
Data Management
Ecology
Geography
Antimicrobials
Applied Mathematics
Simulation and Modeling
Statistics
Drugs
Genome project
Genomics
Anti-Bacterial Agents
Breast Feeding
Computational Theory and Mathematics
Medical Microbiology
Modeling and Simulation
Physical Sciences
Female
Algorithms
Research Article
Computer and Information Sciences
QH301-705.5
Microbial Genomics
Machine learning
Research and Analysis Methods
Data type
Microbiology
03 medical and health sciences
Cellular and Molecular Neuroscience
Machine Learning Algorithms
Artificial Intelligence
Microbial Control
Genetics
Humans
Microbiome
Statistical Methods
Molecular Biology
Ecology, Evolution, Behavior and Systematics
Taxonomy
Pharmacology
business.industry
Biology and Life Sciences
Infant
Models, Theoretical
Country of origin
030104 developmental biology
Metagenomics
Women's Health
Artificial intelligence
Neonatology
business
computer
Breast feeding
030217 neurology & neurosurgery
Mathematics
Forecasting
Subjects
Details
- Language :
- English
- ISSN :
- 15537358 and 1553734X
- Volume :
- 16
- Issue :
- 5
- Database :
- OpenAIRE
- Journal :
- PLoS Computational Biology
- Accession number :
- edsair.doi.dedup.....320d7e429de32323d1613e90dc6ffdec