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 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.

[1]  Francesco Asnicar,et al.  QIIME 2: Reproducible, interactive, scalable, and extensible microbiome data science , 2018 .

[2]  Cesare Furlanello,et al.  Phylogenetic convolutional neural networks in metagenomics , 2017, BMC Bioinformatics.

[3]  Aly A. Khan,et al.  Lactobacillus rhamnosus GG-supplemented formula expands butyrate-producing bacterial strains in food allergic infants , 2015, The ISME Journal.

[4]  Derek Reiman,et al.  Using convolutional neural networks to explore the microbiome , 2017, 2017 39th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[5]  Gregory Ditzler,et al.  Multi-Layer and Recursive Neural Networks for Metagenomic Classification , 2015, IEEE Transactions on NanoBioscience.

[6]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[7]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[8]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[9]  L. T. Angenent,et al.  Succession of microbial consortia in the developing infant gut microbiome , 2010, Proceedings of the National Academy of Sciences.

[10]  C. Mackay,et al.  Dietary Fiber and Bacterial SCFA Enhance Oral Tolerance and Protect against Food Allergy through Diverse Cellular Pathways. , 2016, Cell reports.

[11]  Derek Reiman,et al.  PopPhy-CNN: A Phylogenetic Tree Embedded Architecture for Convolution Neural Networks for Metagenomic Data , 2018, bioRxiv.

[12]  Richard Hans Robert Hahnloser,et al.  Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit , 2000, Nature.

[13]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[14]  G. Weinstock,et al.  A prospective microbiome‐wide association study of food sensitization and food allergy in early childhood , 2018, Allergy.

[15]  Fuhui Long,et al.  Feature selection based on mutual information criteria of max-dependency, max-relevance, and min-redundancy , 2003, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[16]  Geoffrey E. Hinton,et al.  Speech recognition with deep recurrent neural networks , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[17]  Yun‐Cheol Na,et al.  Faecalibacterium prausnitzii subspecies-level dysbiosis in the human gut microbiome underlying atopic dermatitis. , 2016, The Journal of allergy and clinical immunology.

[18]  Razvan Pascanu,et al.  On the difficulty of training recurrent neural networks , 2012, ICML.

[19]  Li Yuan,et al.  Altered Fecal Microbiota Composition Associated with Food Allergy in Infants , 2014, Applied and Environmental Microbiology.

[20]  Tommi Vatanen,et al.  Variation in Microbiome LPS Immunogenicity Contributes to Autoimmunity in Humans , 2016, Cell.

[21]  G. Gerber,et al.  A microbiota signature associated with experimental food allergy promotes allergic sensitization and anaphylaxis. , 2013, The Journal of allergy and clinical immunology.

[22]  Huaiyu Zhu.  On Information and Sufficiency , 1997 .

[23]  R. Tibshirani.  Regression Shrinkage and Selection via the Lasso , 1996 .

[24]  Mihai Pop,et al.  Longitudinal analysis of the lung microbiota of cynomolgous macaques during long-term SHIV infection , 2016, Microbiome.

[25]  Alexander Grishin,et al.  Early-life gut microbiome composition and milk allergy resolution. , 2016, The Journal of allergy and clinical immunology.

[26]  H. Akaike.  Fitting autoregressive models for prediction , 1969 .

[27]  D. Antonopoulos,et al.  Commensal bacteria protect against food allergen sensitization , 2014, Proceedings of the National Academy of Sciences.

[28]  Jianxin Shi,et al.  Allergy associations with the adult fecal microbiota: Analysis of the American Gut Project , 2015, EBioMedicine.

[29]  Botanic Gardens,et al.  In Early Childhood , 2017 .

[30]  Patricia W. Finn,et al.  WEVOTE: Weighted Voting Taxonomic Identification Method of Microbial Sequences , 2016, bioRxiv.

[31]  P. Finn,et al.  Microbiome: Allergic Diseases of Childhood , 2018 .

[32]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[33]  Tommi Vatanen,et al.  The dynamics of the human infant gut microbiome in development and in progression toward type 1 diabetes. , 2015, Cell host & microbe.

[34]  Alex Bateman,et al.  An introduction to hidden Markov models. , 2007, Current protocols in bioinformatics.

[35]  A. Knulst,et al.  Fecal Microbiome and Food Allergy in Pediatric Atopic Dermatitis: A Cross-Sectional Pilot Study , 2018, International Archives of Allergy and Immunology.

[36]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[37]  Luke R. Thompson,et al.  Species-level functional profiling of metagenomes and metatranscriptomes , 2018, Nature Methods.

[38]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[39]  R. Knight,et al.  Moving pictures of the human microbiome , 2011, Genome Biology.

[40]  Duy Tin Truong,et al.  MetaPhlAn2 for enhanced metagenomic taxonomic profiling , 2015, Nature Methods.

[41]  R. Kumar,et al.  The Prevalence, Severity, and Distribution of Childhood Food Allergy in the United States , 2011, Pediatrics.

[42]  Ahmed A. Metwally,et al.  Analysis of the microbiome: Advantages of whole genome shotgun versus 16S amplicon sequencing. , 2016, Biochemical and biophysical research communications.

[43]  David Vandyke,et al.  Semantically Conditioned LSTM-based Natural Language Generation for Spoken Dialogue Systems , 2015, EMNLP.

[44]  Yang Dai,et al.  MetaLonDA: a flexible R package for identifying time intervals of differentially abundant features in metagenomic longitudinal studies , 2018, Microbiome.

[45]  B. Roe,et al.  A core gut microbiome in obese and lean twins , 2008, Nature.