README.md

In syma-research/microTensor: Tensor Decomposition for Longitudinal Microbiomes

microTensor: Tensor Decomposition for Longitudinal Microbiome Data

This repository contains the method implementation and analysis scripts used for the paper "A Tensor Decomposition Model For Longitudinal Microbiome Studies", by Siyuan Ma and Hongzhe Li.

This is the main branch, which is used for the software and analysis used in the paper! If you are interested in the R package simply implementing the microTensor method, please switch to the "package" repo.

To use this repo to reproduce our paper:

1. Download the repository to a local directory.

2. In R, navigate your working directory to where the repo is stored.

3. Make sure you are under the main branch. In R, install the package with devtools::install()

   • This installs the version of microTensor that is necessary to reproduce this analyses performed in the paper.

4. rmd/ has the scripts for analyses performed in the paper.

   • 0_Format_Data.Rmd formats the abundance tables and metadata from the FARMM and DIABIMMUNE cohorts, such that they are ready for downstreama analyses. Since the formatted .RData files are also included in this repo, you can actually skip this file.

   • 1_Simulation_TrueMod.Rmd, 2_Simulation_MisMod.Rmd, and 3_Simulation_SparseDOSSA.Rmd are scripts for the three simulation studies conducted in the paper. These contain large-scale simulation computation, and are meant to be run on a high-performance computing server. It uses the batchtools R package to schedule jobs, which you should configure according to your HPC workload manager.

   • 4_FARMM.Rmd and 5_DIABIMMUNE.Rmd are the scripts used to analyze data and visualize results for the two-real world studies. If you are interested in quick, reproducible analyses from the paper, these are the best targets, as they can be run on a local computer.

5. In addition to the packages necessary for the software implementation, as specified in DESCRIPTION, the following R packages are also necessary for the analyses scripts:

   • tidyverse, for various data table manipulation and visualization utilities.
   • As mentioned, batchtools was used to submit and inspect jobs on HPC clusters for the simulation studies.
   • SparseDOSSA which is the simulation software used for simulation study 3.
   • lmerTest was used for testing on the correlated loadings identified by PCA.

Contact: Siyuan Ma [e-mail]

syma-research/microTensor documentation built on July 1, 2022, 6:30 a.m.