AGENT-P: The Association of Gene Expression to Neuroimaging Traits Pipeline #71
Description
Title
AGENT-P
Leaders
Nhung Hoang (@nhunghoang)
Principal Investigator: Mika Rubinov, PhD
Collaborators
Wenrong Wu
Project description
This project aims to develop a standardized workflow for comparative transcriptome-wide association studies (TWAS) across multi-modal brain phenotypes (e.g., neuroimaging, clinical, morphology, etc). The TWAS framework is an established method in genetics literature for studying the underlying genetic influences on complex traits. This approach is relatively new to the neuroimaging field, but has already revealed many gene-trait associations for brain measures that have highly regional and individual variability. The relative novelty of TWAS in neuroimaging studies means there is a lack of standards and guidelines for producing robust and reproducible results on image-derived traits. Furthermore, in addition to the different estimation methods available, distinct genotype and gene expression preprocessing steps can have varying effects on results and may be unfamiliar to neuroimaging researchers.
We have developed AGENT-P, a Python package geared towards neuroimaging scientists for performing TWAS, and are looking for a team of students to help us improve the latest version.
Link to project repository/sources
https://github.com/nhunghoang/dev_agentp
Concerete goals with specific tasks for Brainhack Vanderbilt 2025
Beginner-friendly deliverables:
- Run and document installation steps for the AGENT-P python package and Dockerfile on various operating systems
- Create a Read The Docs webpage for AGENT-P (and clean up the existing documentation)
Intermediate-level deliverables:
- Add input validation steps to any of the major AGENT-P modules (i.e., check that input data is formatted correctly and provide informative error messages)
- Generate synthetic neuroimaging and genotype data that reflects standard file formats and existing datasets (so we can robustly test common and corner cases)
Advanced deliverables:
- Refactor sections of existing code for computational optimization and/or making it easy to add new features in the future (e.g., remove hardcoded things)
- Devise a strategy for keeping AGENT-P updated with respect to its main dependencies (some of which are active Git repos as well)
- Incorporate existing data visualization scripts into their corresponding AGENT-P modules
Good first issues
Walk through the existing README (which contains installation and running steps) and try to get AGENT-P to work on your machine. Example data and script are already available in the Git repo.
Skills
Necessary:
- Python
Helpful:
- Unix
- Docker
- R
- object-oriented programming
- parallel processing
Bonus:
- any experience with bioinformatics tools (e.g., plink2)
Onboarding documentation
See README of the project Git repo.
What will participants learn?
Example skills you'll learn:
- how a python package is developed
- what brain genomics research looks like
- considerations for generating synthetic data
- object-oriented programming
- parallel processing
- refactorization
- input validation
- modularization
- multi-modal data analysis (specifically for neuroimaging and genetics data)
- data visualization
- how to write user-friendly documentation
Public data to use
Example data is available on the project Git repo.
Number of collaborators
4+
Credit to collaborators
Project contributors will be acknowledged in the eventual journal article, as well as all online sources of AGENT-P (pypi, Git repo, etc). Co-authorship is available for substantial contributions.
Image
Project Summary
AGENT-P is a Python package that supports and streamlines transcriptome-wide association studies for multi-modal brain phenotypes.
Type
method_development, documentation, pipeline_development
Development status
2_releases_existing
Topic
statistical_modelling, data_visualisation, systems_neuroscience
Tools
other
Programming language
Python, R, containerization, documentation, unix_command_line, shell_scripting
Modalities
not_applicable, other
Git skills
0_no_git_skills
Anything else?
No response
Things to do after the project is submitted and ready to review.
- Add a comment below the main post of your issue saying:
Hi @brainhack-vandy/project-monitors my project is ready!