https://github.com/jcaperella29/methylation_shiny

An interactive R Shiny web app for DNA methylation analysis with support for DE, pathway enrichment, power analysis, and machine learning — deployable on HPC via Singularity/Apptainer.
https://github.com/jcaperella29/methylation_shiny

apptainer bioconductor bioinformatics data-visualization differential-expression epigenetics genomics machine-learning methylation pathway-enrichment-analysis pca r random-forest shiny singularity umap

Last synced: 10 months ago
JSON representation

An interactive R Shiny web app for DNA methylation analysis with support for DE, pathway enrichment, power analysis, and machine learning — deployable on HPC via Singularity/Apptainer.

• Host: GitHub
• URL: https://github.com/jcaperella29/methylation_shiny
• Owner: jcaperella29
• Created: 2025-05-12T15:16:01.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2025-06-10T20:43:12.000Z (12 months ago)
• Last Synced: 2025-06-10T21:38:16.499Z (12 months ago)
• Topics: apptainer, bioconductor, bioinformatics, data-visualization, differential-expression, epigenetics, genomics, machine-learning, methylation, pathway-enrichment-analysis, pca, r, random-forest, shiny, singularity, umap
• Language: R
• Homepage:
• Size: 269 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.txt

Awesome Lists containing this project

README

          
DNA Methylation Analysis Shiny App

==================================

Overview

This Shiny app allows users to upload DNA methylation data and perform a complete downstream analysis, including visualization, differential methylation, pathway enrichment, power analysis, and classification via Random Forest.

File Requirements

- Input must be a .csv file.

- The file should include:

•	A 'Gene' column (probe/gene identifiers)

•	A 'Region' column (e.g., Island, Shore, Shelf)

•	Multiple sample columns (e.g., Tumor1, Tumor2, Normal1, etc.)

Usage Instructions

1. Upload your methylation data using the file input.

2. Select the sample groups (detected automatically).

3. Use the buttons in the sidebar to:

    - Generate boxplots and heatmaps

    - Run differential methylation (DE) analysis

    - Perform pathway enrichment (via enrichR)

    - Conduct power analysis

    - Visualize sample clustering (PCA / UMAP)

    - Train and evaluate a Random Forest classifier

Random Forest Tab

- Trains a classifier using probes with adjusted p-value ≤ 0.05 from DE results.

- Includes three subtabs:

•	Prediction: actual vs predicted group labels

•	Metrics: accuracy, sensitivity, specificity, AUC

•	Feature Importance: top features ranked by importance

- Each subtab includes a download button.

Tips

- Choose exactly two groups for DE, power, and Random Forest.

- Use consistent sample name prefixes (e.g., Tumor1, Tumor2).

- All plots are interactive (via Plotly) where applicable.

- Download buttons are provided for results in each section.

Contact

jcaperella@gmail.com

Note

If you encounter issues, double-check your input formatting, or check for NA values in key columns.