🧪Fermentation Explorer: A Scientific App with Features for the Entire Shiny Community - Shiny Contest Submission

tjhackmann · September 7, 2024, 10:20pm

Fermentation Explorer: A Scientific App with Features for the Entire Shiny Community

Authors: Timothy Hackmann

Abstract: Fermentation Explorer is an app used by microbiologists, but it has features useful to the entire Shiny community. Originally described in the journal Science Advances, the app allows microbiologists to explore microbes that carry out fermentation. For the Shiny community, it presents several new features for building apps, such as new types of interactive plots and UI elements.

Full Description:
Key Features

Original version published in a scientific journal
Interactive plots
- 3D network graphs
  - Built using core functions of Plotly
  - Accepts igraph objects as input
- Phylogenetic trees
  - Built using core functions of Plotly
  - Accepts ggtree objects as input
- Treemaps and heatmaps
  - Extends Plotly’s built-in treemap and heatmap functionality
New UI elements for Shiny
- Navigation Buttons
  - Large buttons with icon, title, and subtitle (ideal for a home page)
- Validation Modal
  - Reports an error message if an input (e.g., uploaded file) is invalid
  - Built on shiny::validate() and shiny::showModal()
- File Input with Modal
  - Built on fileInput() and includes a link to a modal for more actions or information
Improved predictive tools for data science
- Metabolic networks and flux-balance analysis
  - Built by extending base functions of fbar
- Predictions from taxonomy
  - Built by implementing basic concept of FAPROTAX in R
- Predictions from machine learning
  - Built using randomForest
Video tutorials

Gallery

3D network graph
Phylogenetic tree

Shiny app: https://timothy-hackmann.shinyapps.io/FermentationExplorer/
Repo: GitHub - thackmann/FermentationExplorer: 🧪A resource for showing the incredible diversity of fermentative metabolism

Thumbnail:
thumbnail_small

Full image:

Strongestdog2016 · September 13, 2024, 9:19pm

This is really cool, Dr. Hackmann! Will users be able to train their own random forest models in the future? Look forward to seeing what you do with this next. Go UC Davis!

tjhackmann · September 14, 2024, 6:21pm

Go UC Davis! Yes, users will be able to train their random forest models in the future. The main challenge has been in building the UI for it, but it is coming.