Enrichard

Enrichard, also known as FLASH (Fast Loci Annotation of Significant enricHments), is a powerful and user-friendly web application designed for CpG site enrichment analysis. It allows researchers to identify over-represented biological traits, pathways, or genomic features within a given set of CpG probe IDs. The tool is built with Python, Streamlit, and a high-performance data processing backend, providing a seamless and interactive experience for complex epigenetic analyses.

The project is built with a modular architecture, making it easy to extend and customize. New analysis modules can be added with minimal effort. To ensure maximum performance and efficiency, the data processing pipeline is heavily optimized with Polars.

🚀 Features

• Three Analisys Pipelines:

  • Single CpG list enrichment: Analyze a single list of CpG sites to find enriched biological terms.

  • Two CpG lists Comparison: Compare two distinct sets of CpG sites to identify common and unique enriched traits.

  • Multi-Sample Group Comparison: Perform a meta-analysis on entire cohorts to identify robust biological themes that are consistently enriched across many samples compared to random controls.

• High Performance: The analysis core is designed for speed, leveraging parallel processing, pre-computation, and the Polars library to deliver results in minutes, not hours.

• Custom Backgrounds: Use predefined genomic backgrounds (HM450, EPIC, EPIC+, EPICv2) or upload your own.

• Multiple Correction Methods: Choose from a variety of statistical correction methods, including Bonferroni and FDR (Benjamini-Hochberg), to control for false positives.

• Interactive Visualizations: Explore your results through dynamic dot plots, heatmaps, and dumbbell plots, making it easy to interpret complex data.

📸 Screenshots

Here's a glimpse of the Enrichard interface and its interactive visualizations.

Main Interface Dot Plot Visualization

Heatmap Visualization

Dumbbell Plot for Comparisons

⚙️ How It Works

Enrichard's backend is built on a robust and modular pipeline that performs the following steps:

1. Data Input: Users can upload a file containing a list of CpG probe IDs or use one of the provided example datasets.
2. Enrichment Analysis: The application uses Fisher's exact test to determine the statistical significance of the enrichment of each biological trait in the provided sample.
3. Multiple Testing Correction: To account for the large number of tests performed, the raw p-values are adjusted using a selected multiple testing correction method.
4. Results Visualization: The final results, including the enriched traits, p-values, and other relevant statistics, are presented in a clear and interactive table.

For two-sample comparisons, the pipeline runs the enrichment analysis for both samples in parallel and then identifies the common traits, providing a side-by-side comparison of the results.

🛠️ Setup and Installation

To run Enrichard locally, you will need to have Docker, Docker Compose, and Git LFS installed.

Some modules, use large pre-computed data files that are stored using Git Large File Storage (LFS).

# Install Git LFS
sudo pacman -S --needed git-lfs

# Install for your user account
git lfs install

Clone the repository and ensure you pull the LFS files.

git clone https://github.com/Serotav/Enrichard
cd Enrichard
git lfs pull
docker compose up --build -d