Takeaway: Tractor Nextflow makes ancestry-aware GWAS harmonized, reproducible, and scalable, helping researchers include admixed individuals and uncover genetic signals missed by traditional methods.

TractorWorkflow/resources at main · Atkinson-Lab/TractorWorkflow Nextflow pipeline for Tractor GWAS with customizable phasing, local ancestry inference (RFMix2, GNomix, FLARE), and regression modules. - Atkinson-Lab/TractorWorkflow

We also provide pre-curated LAI reference panels for 2-way AFR–EUR and 3-way AFR–EUR–AMR from TGP-HGDP reference dataset.

Check: github.com/Atkinson-Lab...

Tractor Nextflow Workflow solves this:
* Automates all steps from QC’d VCF to Tractor GWAS
* Modular & scalable
* Supports multiple tools: SHAPEIT5, RFMix2, FLARE, GNomix
* Can run anywhere: local, HPC, cloud

However, running Tractor is a multi-step process:
* Module 1: Accurate phasing
* Module 2: Local ancestry inference
* Module 3: Tractor GWAS

This means multiple tools, scripts and reference panels to decide from.

Solution: Tractor GWAS
* Models local ancestry for robust association testing
* Produces ancestry-specific summary statistics
* Detects ancestry-enriched hits missed by standard GWAS

Documentation: atkinson-lab.github.io/TractorWorkf...

Problem: Admixed individuals are often excluded due to population stratification, even though they make up a substantial portion of modern biobanks. This could lead to missed genetic discoveries.

Tractor Workflow Pipeline: A Scalable Nextflow Framework for Local Ancestry-Aware Genome-Wide Association Studies The routine exclusion of admixed individuals from traditional Genome-Wide Association Studies (GWAS) due to concerns about spurious associations has hindered genetic analyses involving multiple ancest...

🚨 New preprint! Traditional GWAS often exclude admixed individuals, missing ancestry-enriched genetic signals. We built Tractor Nextflow Workflow - a modular, scalable pipeline that automates phasing, LAI, and Tractor GWAS, alongside comprehensive documentation.
Preprint: doi.org/10.1101/2025...