H5adToGCT

A tool for converting h5ad files to GCT format.

Description

H5adToGCT is a Python-based tool that converts AnnData (h5ad) files to Gene Cluster Text (GCT) format. This tool is particularly useful for single-cell RNA-seq data analysis workflows.

Features

• Convert h5ad files to GCT format
• Include metadata fields in the output
• Specify which layer to use for expression values
• Verbose output option for detailed processing information

Dependencies

The Docker image includes the following dependencies:

• Python 3.9
• scanpy
• pandas
• anndata
• h5py
• humanfriendly

Docker Usage

Building the Docker Image

To build the Docker image:

docker build -t h5adtogct .

This will create a Docker image based on Python 3.9-slim and install all necessary dependencies.

Running the Docker Container

Basic usage:

docker run -v $(pwd):/data h5adtogct -i /data/your_file.h5ad

The output will be saved as output.gct in the current directory.

Command Line Options

• -i, --input: Input h5ad file (required)
• -m, --metadata: Comma-separated list of metadata fields to include (optional)
• -l, --layer: Layer to use for expression values (default: use .X) (optional)
• -v, --verbose: Increase output verbosity (optional)

Examples

Convert a file with verbose output:

docker run -v $(pwd):/data h5adtogct -i /data/your_file.h5ad -v

Include metadata fields:

docker run -v $(pwd):/data h5adtogct -i /data/your_file.h5ad -m "cell_type,batch"

Use a specific layer:

docker run -v $(pwd):/data h5adtogct -i /data/your_file.h5ad -l "normalized"

Helper Script

A helper script build_and_run.sh is provided to simplify building and running the Docker container:

# Build the Docker image
./build_and_run.sh build

# Run the container
./build_and_run.sh run /path/to/your_file.h5ad -v

# Show help
./build_and_run.sh help

GenePattern Module

This tool is also available as a GenePattern module. The module can be run through the GenePattern web interface or programmatically using the GenePattern API.

License

[Add license information here]