Abstract: Mendelian randomization (MR) is a causal inference method that utilizes genetic variations as instrumental variables to investigate the causal relationships between exposures (various factors) and outcomes (diseases or phenotypes). This article discusses the fundamental assumptions of MR methods in biomedical research, common designs, key challenges, optimizations and the future prospects of MR. The article introduces the core assumptions of MR methods and how to select common MR design types in biomedical research. Furthermore, the article discusses the key challenges currently faced by MR studies and their solutions, such as how to select instrumental variables that meet the core assumptions, how to choose causal estimation methods in MR, and how to interpret MR results biologically. In addition, the article suggests some directions for future research. As a powerful tool, MR contributes to exploring causal relationships, selecting therapeutic intervention targets, and conducting long-term population-based intervention studies. While MR cannot fully replace randomized controlled trials, it has a wide  application in biomedical research and clinical practice, helping us to deeply understand the complex relationships between health and disease.

Key words: Mendelian randomization analysis, common designs,  challenges