Identification of Multi-tissue Protein Quantitative Trait Loci and Causal Inference of Protein Effects in Neurological and Other Complex Human Diseases via Mendelian Randomization

Alzheimer disease (AD) is the most common neurodegenerative disease, affecting 6.2 million people mostly aged 65 years or older in the United States as of early 2021. AD has been widely studied and characterized worldwide, but there is still no effective treatment or cure. Even the latest FDA-approved treatment, Aducanumab, cannot stop decline or improve cognition. To develop a truly effective treatments, researchers keep discovering genetic and molecular mechanisms underlying the disease. Protein biomarkers are keys to bridge the mechanisms to disease. Here, I first used a high-throughput proteomic dataset from three tissues (CSF, plasma, and brain) with array-based genotype data to discover protein quantitative trait loci (pQTLs) within each tissue. Next, I replicated most of these findings with publicly available single tissue pQTL datasets. I identified over 10 independent, local pQTLs and pleiotropy regions. I investigated the tissue-specificity and genetic colocalization of pQTLs with QTLs of other molecular traits. Finally, I used Mendelian randomization to infer protein causal effects on neurological and other complex diseases. Taken together, these findings prioritized the proteins that play important role in human disease pathogenesis and helped identify potential biomarkers for AD or other human diseases