Inferring Causal Cell Types Driving Human Disease and Complex Traits - MPG Primer 2023

Intro
Disease-associated cell types can be identified by integrating GWAS with functional genomics data
Colocalization of eQTLs with GWAS variants can implicate disease-critical genes and tissues
This is analogous to the need for fine-mapping GWAS variants
We want a method that can identify the causal tissues among many tagging tissues
Transcriptome-wide association studies (TWAS) perform polygenic colocalization of genes with disease
TWAS association statistics are proportional to the amount of tagged causal effects due to co-regulation
Co-regulation across tissues and genes can be estimated using gene expression prediction models and a reference panel
Visualization of multivariate regression in TCSC
TCSC is powerful, well-calibrated, and unbiased in simulations
TCSC power is modest but can improve by modifying certain parameters
eQTL sample size is an important consideration for real trait analysis
Applying TCSC to real gene expression and trait data
TCSC identifies causal tissue-trait pairs
TCSC performs well when model assumptions are violated
Getting started with TCSC
Data access provided on the TCSC Repo