PT  - JOURNAL ARTICLE
AU  - Ko Abe
AU  - Masaaki Hirayama
AU  - Kinji Ohno
AU  - Teppei Shimamura
TI  - A Latent Allocation Model for the Analysis of Microbial Composition and Disease
AID  - 10.1101/396903
DP  - 2018 Jan 01
TA  - bioRxiv
PG  - 396903
4099  - http://biorxiv.org/content/early/2018/08/21/396903.short
4100  - http://biorxiv.org/content/early/2018/08/21/396903.full
AB  - Background Establishing the relationship between microbiota and specific disease is important but requires appropriate statistical methodology. A specialized feature of microbiome count data is the presence of a large number of zeros, which makes it difficult to analyze in case-control studies. Most existing approaches either add a small number called a pseudo-count or use probability models such as the multinomial and Dirichlet-multinomial distributions to explain the excess zero counts, which may produce unnecessary biases and impose a correlation structure taht is unsuitable for microbiome data.Results The purpose of this article is to develop a new probabilistic model, called BERMUDA (BERnoulli and MUltinomial Distribution-based latent Allocation), to address these problems. BERMUDA enables us to describe the differences in bacteria composition and a certain disease among samples. We also provide a simple and efficient learning procedure for the proposed model using an annealing EM algorithm.Conclusion We illustrate the performance of the proposed method both through both the simulation and real data analysis. BERMUDA is implemented with R and is available from GitHub (https://github.com/abikoushi/Bermuda).