This technology is a statistical method that uses a penalized inverse-variance weighted estimator with closed-form correction to improve causal inference in Mendelian randomization studies.
Current methods for Mendelian randomization (MR), such as the inverse-variance weighted (IVW) estimator, are highly sensitive to weak genetic instruments and horizontal pleiotropy, leading to biased or unreliable causal estimates. These limitations reduce confidence in using MR to prioritize drug targets, validate biomarkers, or guide clinical decision-making, especially for complex traits like cardiometabolic and infectious diseases. There is a critical need for MR approaches that improve accuracy and robustness without sacrificing computational efficiency, enabling reliable causal inference from large-scale genomic data.
This technology is a statistical method that improves the reliability of Mendelian randomization (MR) by incorporating a penalty term into the standard inverse-variance weighted (IVW) estimator to stabilize calculations when genetic instruments are weak. A closed-form correction further reduces bias caused by horizontal pleiotropy, improving the accuracy of causal effect estimates while maintaining computational efficiency. Implemented as an R software package, this method can be integrated into existing MR workflows for large-scale genomic analyses.
This technology has been validated through simulation studies and applied to real-world genetic datasets, including COVID-19 severity and cardiometabolic disease outcomes.
IR CU25374
Licensing Contact: Joan Martinez
