Highly Efficient Imaging Genetics (HEIG)

HEIG is a statistical framework for efficiently conducting joint analysis for large-scale imaging and genetic data. Compared to traditional methods, HEIG reduces computational time and storage burden by over 200 times, significantly boosts statistical power in association analysis, and most importantly, defines the standard to share the voxel-level GWAS summary statistics to the community.

The analysis can be performed by HEIG (will have more in the near future):

• Voxelwise genome-wide association analysis (VGWAS), including effcient GWAS for high-dimensional non-imaging phenotypes
• Voxelwise heritability analysis
• Genetic correlation analysis for pairs of voxels
• Cross-trait genetic correlation between voxels and non-imaging phenotypes

Version

• v1.0.0: initial version of HEIG.
• v1.1.0: support multi-threading computation; many changes in data format; not compatible with v1.0.0.
• v1.2.0: support LDR GWAS; provide more data mangement options; fix bugs in v1.1.0.

System Requirements

OS Requirements

HEIG is supported for macOS and Linux. It has been tested on the following systems:

• Red Hat Enterprise Linux 8.9
• MacOS Sonoma 14.5

Python Dependencies

HEIG is implemented in Python 3.11. Specific package dependencies are provided in requirements.

Getting Started

First download the released version, unzip it, and navigate to the extracted folder:

wget -O heig-1.2.0.zip  https://github.com/Zhiwen-Owen-Jiang/heig/archive/refs/tags/v1.2.0.zip
unzip heig-1.2.0.zip
cd heig-1.2.0

Install Miniconda based on your OS, and build the environment and install all dependencies for HEIG. This step may take ~5 minutes.

conda env create --file environment.yml
conda activate heig

Or you can create a new environment and install dependencies through pip

conda create --name heig python=3.11
conda activate heig
pip install -r requirements.txt

Since version v1.2.0, hail has been a dependency for conducting GWAS analysis. If you fail to install hail, please contact the hail team.

How to use HEIG

We provided detailed tutorial for using HEIG. The example data used in the tutorial for v1.2.0 can be downloaded here. Common issues are described in the FAQ.

If that does not work, email Owen Jiang [email protected] or [email protected].

Citation

TBD.

Licence

This project is licensed under GNU GPL v3.

Authors

Zhiwen (Owen) Jiang and Hongtu Zhu (University of North Carolina at Chapel Hill)