Utilizing longitudinal microbiome taxonomic profiles to predict food allergy via Long Short-Term Memory networks

Food allergy is usually difficult to diagnose in early life, and the inability to diagnose patients with atopic diseases at an early age may lead to severe complications. Numerous studies have suggested an association between the infant gut microbiome and development of allergy. In this work, we inv...

Full description

Saved in:

Bibliographic Details
Published in:	PLoS computational biology 2019-02, Vol.15 (2), p.e1006693-e1006693
Main Authors:	Metwally, Ahmed A, Yu, Philip S, Reiman, Derek, Dai, Yang, Finn, Patricia W, Perkins, David L
Format:	Article
Language:	eng
Subjects:	Allergies Allergy Artificial intelligence Artificial neural networks Atopy Bioengineering Biology and Life Sciences Classification - methods Complications Computer and Information Sciences Computer science Deep learning Dermatitis Diagnosis Energy management Feature extraction Finn, David Food Food allergies Food allergy Food hypersensitivity Food Hypersensitivity - genetics Forecasting - methods Funding Humans Immunology International conferences Intestinal microflora Learning algorithms Long short-term memory Longitudinal Studies Machine Learning Markov chains Markov processes Medical research Medicine Medicine and Health Sciences Medicine, Experimental Memory, Long-Term - physiology Memory, Short-Term - physiology Methods Microbiomes Microbiota Microbiota (Symbiotic organisms) Microorganisms Multilayers Natural language processing Neural networks Neural Networks, Computer Phylogenetics Physical sciences Redundancy Representations Retirement benefits Short term memory Social Sciences Supervision Support Vector Machine Support vector machines Taxonomy Training Work capacity
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Food allergy is usually difficult to diagnose in early life, and the inability to diagnose patients with atopic diseases at an early age may lead to severe complications. Numerous studies have suggested an association between the infant gut microbiome and development of allergy. In this work, we investigated the capacity of Long Short-Term Memory (LSTM) networks to predict food allergies in early life (0-3 years) from subjects' longitudinal gut microbiome profiles. Using the DIABIMMUNE dataset, we show an increase in predictive power using our model compared to Hidden Markov Model, Multi-Layer Perceptron Neural Network, Support Vector Machine, Random Forest, and LASSO regression. We further evaluated whether the training of LSTM networks benefits from reduced representations of microbial features. We considered sparse autoencoder for extraction of potential latent representations in addition to standard feature selection procedures based on Minimum Redundancy Maximum Relevance (mRMR) and variance prior to the training of LSTM networks. The comprehensive evaluation reveals that LSTM networks with the mRMR selected features achieve significantly better performance compared to the other tested machine learning models.
ISSN:	1553-7358 1553-734X 1553-7358