UTMOST_fa

UTMOST_fa is an R package for joint gene expression imputation across multiple contexts, incorporating functional annotations. It is particularly tailored for single-cell transcriptome-wide association studies (sc-TWAS), extending the original UTMOST training process with improved modeling of regulatory SNPs.

1. Installation

UTMOST_fa can be installed from this GitHub repository.

# Install the devtools package if you haven't already
# install.packages("devtools")

devtools::install_github("leaffur/UTMOST_fa")

2. Tutorial

A detailed tutorial is under development and will be provided soon.

3. Association tests using external GWAS

UTMOST_fa focuses on building and training gene expression imputation models for TWAS analyses. To apply these trained models on external GWAS summary statistics and identify disease-associated genes, please refer to UTMOST framework.

Note for Python 3 users:

We provide a Python 3-compatible alternative for UTMOST.

To use it:

1. Clone the UTMOST repository

$ git clone https://github.com/Joker-Jerome/UTMOST

2. Download and extract the UTMOST_py3.zip file from the main UTMOST_fa repository folder.

3. Replace the single_tissue_covariance.py file and the metax folder inside the UTMOST directory with the provided Python 3 versions.

4. Reference

• Hu, Y. et al. A statistical framework for cross-tissue transcriptome-wide association analysis. Nature Genetics 51, 568-576 (2019). Link.