Probabilistic Transcriptome-wide Association Analysis (PTWAS)

This repository contains the software implementations and necessary annotation resources for a suite of statistical methods to perform transcriptome-wide association analysis (TWAS). These methods are designed to perform rigorous causal inference connecting genes to complex traits. The statistical models and the key algorithms are described in the manuscript [1].

The repository includes source code, scripts and necessary data to replicate the results described in the manuscript. A detailed tutorial to guide the users through some specific analysis tasks is also included.

For questions/comments regarding to the software package, please contact Xiaoquan (William) Wen (xwen at umich dot edu).

License

Software distributed under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at you r option) any later version. See LICENSE for more details.

Quick Start

Quick start instructions for running PTWAS

Repository directories

• PTWAS_scan: code and tutorial to run PTWAS scan procedure

• PTWAS_est: code and tutorial to run model diagnosis and effect size estimation procedures

• PTWAS_paper: code and downloadable data sources for simulations and real data analysis used in the PTWAS paper

Required software

• DAP: this software package performs Bayesian multi-SNP fine-mapping analysis and generates the probabilistic annotations (at model, signal, and SNP levels) required by PTWAS. Required if you want to use your own eQTL data.

• GAMBIT: this software package implements a fast general burden test (with pre-defined weights). PTWAS requires GAMBIT to perform scan procedure when only summary-level statistics of GWAS, i.e., single-SNP z-scores, are made available.

Available resources

• Pre-built PTWAS annotation file using GTEx (v8) data for 49 tissues: download

Contributors

• Corbin Quick (Harvard University)
• Yuhua Zhang (University of Michigan)
• Xiaoquan Wem (University of Michigan)

Citation

• Chen, Y., Quick, C., Yu, K., Barbeira, A.,The GTEx Consortium, Luca, F., Pique-Regi, R., Im, H.K., Wen, X. Investigating tissue-relevant causal molecular mechanisms of complex traits using probabilistic TWAS analysis. 2020. Genome Biology (21): 232.