Application Deployment Guide

This guide documents standalone deployment files, setup scripts, and Shiny application templates for running and deploying foundrShiny applications located in the inst/shinyApp/ directory.

Data Preliminaries
Directory Overview
Expected Data Directory Layout
Setting Up Data for Shiny App
How to Run the Shiny App

Data Preliminaries

Unless you have already loaded data, you will need to do that with one of the scripts, for either a minimal subset or the full Liver panel. You will need to be on campus or use GlobalProtect VPN.

The processed data for the foundrShiny app are in [RD]/adattie/General/founder_diet_study where [RD] is /research.drive.wisc.edu, which you typically mount as /Volumes for Linux and MacOS and R: for Windows.

On MacOS, it is convenient to establish a symbolic link in your home directory (~) with the one-time command:

ln -s /Volumes/adattie/General/founder_diet_study ~/founder_diet_study

Raw data are in ~/founder_diet_study/RawData and harmonized data are in ~/founder_diet_study/HarmonizedData. These are accessed via the setup scripts as input to the foundrShiny app.

Directory Overview

The inst/shinyApp/ deployment directory contains the following components:

File	Description
`app.R`	Primary Shiny application entry point using `foundrShiny::foundrSetup()`.
`setup.R`	Standalone script for loading datasets into the workspace without `foundrSetup()`.
`TraitData.R`	Helper script sourced by `foundrSetup()` to process and filter input RDS files.
`appDetail.R`	Alternative standalone Shiny deployment script with explicit file path configurations.
`intro.md`	Markdown overview displayed in the entry panel of the Shiny app.
`help.md`	User help documentation displayed within the interactive panels.
`foundrDAG.Rmd`	Directed Acyclic Graph (DAG) documentation file for trait module structures.

Data Prerequisites & Directory Structure

To run the Shiny application, your data directory (dirpath) must contain pre-processed .rds data objects generated by the foundr package pipeline.

Expected Data Directory Layout

dirpath/
├── liverData.rds        # Trait measurements (or traitData.rds)
├── liverSignal.rds      # Statistical signal data (or traitSignal.rds)
├── liverStats.rds       # Differential analysis statistics (or traitStats.rds)
├── traitModule.rds      # Module definitions / clustering (optional)
├── datasets.rds         # Metadata table mapping dataset keys to titles
└── AppSetup/
    └── Liver/
        └── help.md      # (Optional) Instance-specific help file

[!NOTE] The prefix (liver or trait) corresponds to the data_instance parameter passed to foundrSetup().

Setting Up Data for Shiny App

You must have key data files loaded into the R workspace prior to running app.R (or foundrShinyApp()). These can be loaded via foundrSetup() or using the setup.R script.

These create five key R objects in the global environment:

traitData: Data frame containing raw/normalized trait values, strain, sex, diet, and dataset metadata.
traitSignal: Data frame of ANOVA / statistical model summaries per trait.
traitStats: Data frame of pairwise contrasts, effect sizes, and p-values.
traitModule: (Optional) Data frame mapping traits to co-expression/functional modules.
customSettings: Named list of app settings (help file path, condition name, group name, dataset lookup table).

Automated Setup via `foundrSetup()` (Recommended)

In app.R, configure foundrSetup() with your data location and desired data subset:

library(foundrShiny)

# 1. Initialize runtime data objects (with subset of data)
foundrShiny::foundrSetup(
  data_subset   = c("Physio", "MixMod")
)

# 2. Define application title
title <- "Founder Diet Study"

# 3. Construct UI and Server using foundrShiny modules
ui <- shiny::fluidPage(
  shiny::titlePanel(title),
  shiny::sidebarLayout(
    shiny::sidebarPanel(
      foundrShiny::foundrInput("foundr")
    ),
    shiny::mainPanel(
      foundrShiny::foundrOutput("foundr")
    )
  )
)

server <- function(input, output, session) {
  foundrShiny::foundrServer(
    "foundr",
    traitData, traitSignal, traitStats,
    customSettings, traitModule
  )
}

shiny::shinyApp(ui, server)

Manual Data Loading via `setup.R`

If you are running in a custom environment or debugging without foundrSetup(), you can define variables prior to sourcing setup.R:

small <- FALSE  # Set to FALSE for full dataset (defaults to TRUE if unset)
source("inst/shinyApp/setup.R")

If small = TRUE, only the first 2 datasets and the first 2 traits per dataset are included.

How to Run the App

From RStudio / R Console

Launch directly from the repository root:

# Run directly using shiny::runApp
shiny::runApp("inst/shinyApp")

Or open inst/shinyApp/app.R in RStudio and click Run App.

From Command Line

Rscript -e "shiny::runApp('inst/shinyApp', port = 3838, host = '0.0.0.0')"

Application Deployment Guide