Introduction

Ad_PX_pipe is a pipeline for performing genome-wide association studies and imputed transcriptome-wide association studies in admixed populations, finding independent and eQTL colocalized gene signals, and mapping admixture using local ancestry estimations. This wiki is a supplement to Ad_PX_Pipe that better explains what each step and each script does, as well as giving more detail on the significance behind each step. It is expected that all scripts are run from the same directory. It is assumed that the user knows the general purpose and execution of a genome-wide association study and PrediXcan, and I will give a brief introduction into colocalization, backward elimination modeling, and local ancestry mapping. However, these topics are much further better explained here, here, and here in their introductions. You can begin by cloning the repository (git clone https://github.com/aandaleon/Ad_PX_pipe.git) and following the steps in the README. If you use this pipeline, please cite Genetically regulated gene expression underlies lipid traits in Hispanic cohorts.

To know why certain steps are performed, refer to the walkthrough. For information on specific scripts, refer to the scripts page, and for information on specific data formats, refer to the data formats page.

Sample cohort

The test data we’re going to use is a subset of 1000 Genomes Project AMR, a cohort of 320 individuals from the Americas. Individuals include Mexican Ancestry from Los Angeles, USA (MXL); Puerto Ricans from Puerto Rico (PUR); Colombians from Medellin, Colombia (CLM); and Peruvians from Lima, Peru (PEL). These correspond with my study of the Hispanic Community Health Study/Study of Latinos, as these individuals also have multi-continental ancestry. These data are subset to only include 100,000 SNPs for speed.

Principal component plot of AMR