KIBREED_public

A genomic prediction project for plant breeding applications using both statistical models and deep learning approaches.

Prerequisites

• R: Version 4.0.5
• Python: Version 3.8.11

Environment Setup

1. Verify Language Versions

Ensure R 4.0.5 and Python 3.8.11 are installed and properly configured in your PATH (likely by adding their paths to .bashrc).

./check_Langs.sh

2. Python Environment Setup

Create a virtual environment named py_env and install all required Python packages from pyenv.lock:

./setup_py.sh

Activate the environment:

source /path/to/KIBREED_public/py_env/bin/activate

3. R Environment Setup

Restore the R package environment using renv and the renv.lock file:

Rscript setup.R

Once completed, your R environment will be automatically activated when you start R. If that does not happen check the .Rprofile file.

Setup Notes

• Run all commands from the project root directory
• If you encounter version mismatch errors, ensure the correct versions are installed and available in your $PATH
• The input data is available at FigShare. Download those files and put them at /path/to/KIBREED_public/data

Documentation

Guide	Description
GBLUP-based models	Traditional genomic prediction using R/BGLR
CNN-based models	Deep learning genomic prediction with TensorFlow

Data Requirements

Both R and Python scripts expect specific data files in the data/ directory. See the individual analysis guides for detailed data format requirements.

License

This project is licensed under the MIT License. See the LICENSE file for full details.