How ldanilova/fest works
Project Type and Purpose
• This is a single‐page R Shiny application (“FEST”) for analyzing TCR sequencing data from peptide-stimulated cultures.
• It provides an interactive web UI (app.r) that talks to server logic (app.r + manafest_shiny_functions.r) to process input data, run analyses, and render plots/tables.
File Structure Analysis
• app.r – Main Shiny entry point; defines UI and server function.
• manafest_shiny_functions.r – Library of data-processing and analysis functions used by the server.
• userManual.html – Static user guide displayed via the app or provided separately.
• README.md, .gitignore – Documentation and version-control exclusions.
Key Architectural Components
a. Shiny UI Layer
– Defines page layout, input controls (file upload, parameter selectors), output placeholders (plots, tables).
b. Shiny Server Layer
– Reactives and observers that respond to UI inputs, call analysis functions, manage data flow, and push results back to the UI.
c. Analysis Functions Module
– Encapsulated routines in manafest_shiny_functions.r (data cleaning, statistical tests, clonotype expansion analysis).
d. Data Inputs/Outputs
– Input: uploaded TCR sequencing files (CSV/TSV), user parameters.
– Output: dynamic plots (ggplot2?), tables (DT?), downloadable reports.
e. Documentation Component
– userManual.html is served as static content or via an iframe/tab in the UI.
f. External Dependencies
– R packages: shiny, data.table/dplyr, ggplot2, htmltools, etc.
– Web server hosting (e.g., Shiny Server or RStudio Connect).
Relationships and Data Flows
• User interacts with UI controls →
• UI triggers reactive expressions in Server →
• Server invokes functions from manafest_shiny_functions.r →
• Functions read/transform data, run analyses →
• Results (tables, plots) are returned to Server →
• Server pushes results back to UI outputs →
• User views/downloads results.
• Optional: static link to userManual.html.
Architectural Patterns and Design Principles
• Reactive Programming (Shiny): clear separation between UI definition and server logic.
• Modularization: isolate complex algorithms in a separate functions file for maintainability and testability.
• Single‐process model: UI and server run in one R process per user session.