In Nature Genetics, investigators from Harvard T.H. Chan School of Public Health describe a tissue co-regulation score regression (TCSC) they developed that disentangles causal tissues from tagging tissues when analyzing genome-wide association studies and gene expression data. TCSC differs from previous methods as it jointly models contributions of heritability from each tissue to disentangle causal tissues from tagging tissues. They applied TCSC to 78 diseases or traits and gene expression prediction models for 48 Genotype-Tissue Expression project tissues and identified 21 causal tissue-trait pairs. Some findings included biologically plausible new tissue-trait pairs, including associations between the aortic artery and glaucoma, the left ventricle of the heart with platelet count, and the brain hippocampus with ADHD. Meanwhile, they also identified biologically plausible, new tissue-trait covariance pairs, including a positive contribution of the brain substantia nigra and a negative contribution of the pancreas to the covariance of BMI and RBC count, respectively.
New Method to Distinguish Causal, Tagging Tissues
Aug 15, 2023