Powerful and robust inference of complex phenotypes' causal genes with dependent expression quantitative loci by a median-based Mendelian randomization

The effective-median-based Mendelian randomization (EMIC) framework improves causal gene inference for complex phenotypes using GWAS summary statistics.

• EMIC addresses high false-positive rates commonly seen in existing Mendelian randomization methods.
• The method utilizes multiple cis-expression quantitative trait loci (eQTLs) to enhance causal gene inference accuracy.
• In simulated datasets, EMIC demonstrated greater power for identifying causal genes compared to alternative approaches.
• EMIC successfully rediscovered known causal genes associated with complex phenotypes like schizophrenia and bipolar disorder.
• The framework also identified new candidate causal genes, expanding the understanding of genetic influences on complex traits.

AI simplified