A shiny app using the R environment to perform microbial analysis of phylogenetic sequencing data. Dame was specifically designed to work directly with output files from QIIME 1 with as minimal file processing as possible.

The current release (v0.1) assesses α- and β-Diversity measurements, and differential expression analyses of sequencing count data. DAME requires the .BIOM file from QIIME and a .CSV file containing the .BIOM sample labels and metadata (experimental grouping data) associated with each sample. This app utilizes the Shiny framework to allow for dynamic and real-time interaction with virtually all aspects of the data workflow. Where possible, table and graphic outputs utilize D3 for a fully interactive experience.

DAME can be downloaded and run locally providing that the latest version of R and shiny are installed. Github launch will install all required packages. Cut and paste the following script into the R console and DAME will launch and open in browser.

library(shiny)
runGitHub("ACNC-DAME", "bdpiccolo")

DAME requires two files to operate:

1. BIOM file - QIIME output file typically found in the folder (OTU) during otu picking method using either of the programs (pick_open_reference_otus.py, pick_closed_reference_otus.py or pick_de_novo_otus.py).

Note: Use the OTU generated file that has taxonomy details (e.g. otu_table_mc3_w_tax.biom). DAME will fail to recognize OTU table without taxonomy details.

2. BIOM Metadata File - .CSV file containing a column with exact sample labels used in QIIMe analysis and experimental groupings. It is recommended to re-purpose the original map file used in QIIME analysis.

3. TRE file (optional) - .TRE file, typically found in the same folder (OTU) from QIIME output.

BIOM files that are generated through other pipelines (which are in JSON format) must be converted to HDF5 format before loading into DAME.

• Convert from biom to txt:

biom convert –i table.biom –o otu_table.txt –to-tsv –header-key taxonomy

• Convert back to biom:

biom convert –i otu_table.txt –o new_ otu_table.biom –to-hdf5 –table-type=”OTU table”
–process-obs-metadata taxonomy

• Utilizes BIOM file to minimize data manipulation steps
• Experimental group selections automatically render
• Allows user to filter data based on meta-data and/or low abundant sequence reads
• Utilizes Shiny for interactive inputs and reactive tables and graphics
• All tables and graphics are interactive

• Summary statistics before and after filters:

  • Sample prevalence
  • Total Reads
  • Total OTUs

• α-diversity statistics by taxonomic levels:

  • Calculates observed, chao1, ACE, Shannon, Simpson, Inverse Simpson, Fisher indices.
  • Calculates 1-way or multifactor ANOVAs based on meta-data.
  • Output tables are rendered with the DT package.
  • Barplots rendered with the highcharter package.*
  • All data (α-diversity calculations and statistics) are downloadable.

• β-diversity statistics by taxonomic levels:

  • Calculates multiple dissimilarity, distance, and tree based parameters.
  • Calculates several ordination methods, including Principal Co-ordinate Analysis, Non-Metric Multidimensional Scaling, and others.
  • Ordination plots rendered with the scatterD3 package.
  • Tables are rendered with the DT package.

• Differential abundance analysis using Negative Binomial Regression by taxonomic levels:

  • Pairwise comparisons of meta-data using DESeq2 workflow.
  • DESeq2 result table rendered with DT package.
  • Boxplots displayed with either Total Reads or Percent Abundance and rendered with the highcharter package.*
  • All results are downloadable.

This web application was built using Shiny by RStudio using open source software. It heavily relies upon functions from phyloseq, vegan, and DESeq2. We highly endorse and encourage visiting the websites associated with these packages.

Created by:

Brian Piccolo
Assistant Professor
Arkansas Nutrition Research Center
University of Arkansas for Medical Sciences

Kartik Shankar
Sree Chintapalli

The following R packages were utilized in no particular order of importance:

shiny
shinyjs
DT
highcharter*
V8
biomformat
ape
pbapply
tibble
reshape2
phyloseq
dplyr
vegan
DESeq2
scatterD3
RColorBrewer

* This app uses Highsoft software with non-commercial packages. Highsoft software product is not free for commercial use.