Skip to content

Analysis of MI

To investigate the reliability of the Latent Causal Variable (LCV)1 causal inference technique, I used LCV to estimate the Genetic Causality Proportion (GCP) of several well-known risk factors on the risk of myocardial infarction (MI). I also included educational attainment as a negative control. The results are below.

upstream_trait posterior_mean_gcp pvalue_gcp_zero_two_sides rho_est rho_se
Triglycerides 0.929622 6.69159e-62 0.306366 0.0430598
LDL 0.838187 3.9004e-34 0.237036 0.0504649
CRP 0.768056 1.30708e-10 0.274214 0.042724
Educational_Attainment 0.040794 0.513409 -0.142967 0.0308109
  • The strong effect of LDL and triglycerides on risk of MI is consistent with their extensively documented status as causal risk factors2.
  • The high GCP for C-reactive Protein (CRP) is interesting. A cursory search suggests that while CRP is a very strong biomarker for inflammation, which is in turn a strong risk for cardiovascular disease, there is controversy about whether CRP itself actually plays a causal role. Thus The high GCP score may indicate a limitation of LCV as a causal inference technique.
  • The low GCP for educational attainment indicates it has successfully served as a negative control.

  1. Luke J O’Connor and Alkes L Price. Distinguishing genetic correlation from causation across 52 diseases and complex traits. Nature Genetics, 50(12):1728–1734, 2018. URL: https://www.nature.com/articles/s41588-018-0255-0. ↩

  2. Daniel Steinberg. The cholesterol wars: the skeptics vs the preponderance of evidence. Elsevier, 2011. URL: https://www.amazon.com/dp/B0085TMWZ4. ↩