GMH microbiome analalysis package

Intro

This package is build for support analyses of microbiome data at The Research Group for Gut, Microbes, and Health at DTU National Food Institute, Denmark.

The package consists of templates for data import, microbiome description, analysis of individual variables, beta diversity analysis, and differential abundance testing. Additionally, some custom functions used in the templates are included.

Content

The package is structured with Rmarkdown templates that together makes up a fundamental microbiome analysis as well as some functions that simplify parts of the analyses.

Templates

As the package is still a work in progress, this is a complete list of all templates that will be included. Not all of the templates have been finalised, but this will be updated when they are added.

• GMH_import
  Input: DADA2 generated phyloseq object and a metadata csv file.
  Description: Import, cleaning, QC, decontamination, and calculates alpha diversities.
  Output: Two phyloseq objects (with different levels of decontamination) and a structured phyloseq object summarizing the process.

• GMH_description
  Input: One phyloseq objects from GMH_import.
  Description: Generates a general description of the microbiomes in the samples. This does not include any statistical analysis.
  Output: A structured .html document with all tables and plots.

• GMH_test_variables (Lacks visualization part)
  Input: One phyloseq objects from GMH_import or a general data file.
  Description: Organise data and structure analyses of predictor and outcome variables, using the statistical methods defined in GMH_test_variables_code.
  Output: Relevant plots and a structured .html document showing the statistical analyses.

• GMH_test_variables_code (Lacks visualization part)
  Input: None
  Description: Guide to identify and apply the correct statistical test differences and correlations between predictor variables (incl. alpha diversity) and outcome variables.
  Output: Code chunks to insert in GMH_test_variables.

• GMH_beta_diversity
  Input: One phyloseq objects from GMH_import.
  Description: Fundamental analysis of beta diversity, identifying differences between groups and correlations with quantitative variables.
  Output: Relevant plots and a structured .html document showing the beta diversity analysis

• GMH_differential_abundance (In progress, only placeholder template)
  Input: One phyloseq objects from GMH_import.
  Description: Identification of differentially expressed taxa. This is structured around the DAtest package and supplemented with plots.
  Output: Relevant plots and a structured .html document showing the differential abundance analysis.

Functions

The package contains the following functions:

• Project initiation
  initiate_GMH_project(): Creates the folder structure used by the script in the the current working directory. It can also create the package’s Rmarkdown templates with generic names.

• Rarefaction and alpha diversity
  calculate_alpha_diversity(): Calculates alpha diversity as the mean of multiple independent rarefactions.
  Rcurve_data(): Wraps around calculate_alpha_diversity() to calculate data for a rarefaction curve plot.
  multiple_rarefy(): Rarefies each sample multiple times and select the best representation for each sample based on within sample betadiversity comparison.

• Filtering
  filter_abundance(): Filters taxa based on average abundance and merges low abundant taxa into “Others”.
  filter_prevalence(): Filters taxa based on prevalence (presence/absence) and merges low prevalence taxa into “Others”.
  filter_rank(): Filters taxa based on ranked average abundance and merges lower ranked taxa into “Others”.
  filter_variance(): Filters taxa based on coefficient of variation (cov) and merges less variating taxa into “Others”. Low abundant taxa more likely to have high cov so I recommend filtering by abundance or prevalence first.

• Taxa cleaning
  clean_taxa(): Replaces NA values in tax_table() with the lowest identified classification for that taxa.
  sort_taxa(): Sort taxa by full taxonomy for plotting (e.i. taxa from the same phylum is plottet next to each other).

Installation

if (!requireNamespace("remotes")) install.packages("remotes")
if (!requireNamespace("BiocManager")) install.packages("BiocManager")

BiocManager::install(version = "3.20")
BiocManager::install("phyloseq")

remotes::install_github("MSMortensen/GMHmicrobiome")

For full functionality of the DAtest package also install the following packages

BiocManager::install(c("DESeq2","limma","edgeR","metagenomeSeq","baySeq","ALDEx2","impute","ANCOMBC"))

install.packages(c("samr","pscl","statmod","mvabund","eulerr","lsmeans"))

How to use

I recommend starting with a clean RStudio project and then follow the steps below:

• Run initiate_GMH_project() to create the necessary folders and generic templates.

• Copy input data to the input folder

• Import data

  • Create a new Rmarkdown file from the GMH_import template
    (File > New File > R Markdown … > From Template > GMH_import).
    > SKIP IF FILES WERE CREATED

  • Modify the template as necessary and when all works, knit the Rmarkdown.

• Describe the microbiome

  • Create a new Rmarkdown file from the GMH_import template
    (File > New File > R Markdown … > From Template > GMH_description).
    > SKIP IF FILES WERE CREATED

  • Modify the template as necessary and when all works, knit the Rmarkdown.

• Run any of the three analysis templates.