MicrobiomeGPS is an application which for exploratory data analysis of microbiome count data, available in both command line (R) and GUI (Shiny) versions. Current modules include:
- Alpha diversity
- Beta diversity
- Differential abundance analysis (taxonomic or functional)
- Predictive modeling
- Community subtype analysis
- OTU network analysis
We include the most robust and powerful statistical methods developed for microbiome data. Specifically, we use linear model/mixed effects modeling for alpha-diversity, MiRKAT/PERMANOVA for beta-diversity, permutation-based FDR control for DA analysis, tree-based random forests for predictive modeling, and SPIEC-EASI for OTU network analysis.
MicrobiomeGPS integrates covariate adjustment in each step and thus can address potential confounding effects due to technical, biological and clinical variables. It can also analyze repeated measurement data by taking into account the within-subject correlation structure. MicrobiomeGPS also allows users to import/export all parameters and download detailed HTML reports to facilitate collaboration and reproducibility.
Starting the MicrobiomeGPS Shiny app is quite simple. In an R session, simply enter:
shiny::runGitHub("SJohnsonMayo/MicrobiomeGPS")
MicrobiomeGPS is also currently deployed on a free shinyapps.io server:
https://sjohnsonmayo.shinyapps.io/github_MicrobiomeGPS/
Watch this space for a manual and vignettes.
MicrobiomeGPS is still under active development. Upcoming features and any bugs can be found on the issues page.