Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2025 Jun 5;112(6):1344-1362.
doi: 10.1016/j.ajhg.2025.04.010. Epub 2025 May 15.

A flexible machine learning Mendelian randomization estimator applied to predict the safety and efficacy of sclerostin inhibition

Marc-André Legault  1 , Jason Hartford  2 , Benoît J Arsenault  3 , Archer Y Yang  4 , Joelle Pineau  5
Affiliations

A flexible machine learning Mendelian randomization estimator applied to predict the safety and efficacy of sclerostin inhibition

Marc-André Legault et al. Am J Hum Genet. .

Abstract

Mendelian randomization (MR) enables the estimation of causal effects while controlling for unmeasured confounding factors. However, traditional MR's reliance on strong parametric assumptions can introduce bias if these are violated. We describe a machine learning MR estimator named quantile instrumental variable (Quantile IV) that achieves a low estimation error in a wide range of plausible MR scenarios. Quantile IV is distinctive in its ability to estimate nonlinear and heterogeneous causal effects and offers a flexible approach for subgroup analysis. Applying quantile IV, we investigate the impact of circulating sclerostin levels on heel bone mineral density, osteoporosis, and cardiovascular outcomes. Employing various MR estimators and colocalization techniques, our analysis reveals that a genetically predicted reduction in sclerostin levels significantly increases heel bone mineral density and reduces the risk of osteoporosis while showing no discernible effect on ischemic cardiovascular diseases. As a second application, we estimated the effect of increases in low-density lipoprotein cholesterol and waist-to-hip ratio on ischemic cardiovascular diseases using this well-known association as a positive control analysis. Quantile IV contributes to the advancement of MR methodology, and the selected applications demonstrate the applicability of our estimator in various MR contexts.

Keywords: Mendelian randomization; UK Biobank; cis-MR; colocalization; genetic epidemiology; instrumental variable; machine learning; proteomics; sclerostin.

PubMed Disclaimer

Conflict of interest statement

Declaration of interests J.H. was an employee of Recursion during the course of this work and has received optional ownership interest in Recursion. B.J.A. is a consultant for Eli Lilly, Silence Therapeutics, Editas Medicine, and Novartis and has received research contracts from Pfizer, Ionis Pharmaceuticals, Eli Lilly, and Silence Therapeutics.

Figures

None
Graphical abstract
Figure 1
Figure 1
Root-mean-square error between the estimated IV regression and the true causal function over a grid spanning 95% of the empirical range of the exposure The boxplot for every estimator represents variability over 200 simulation replicates. The simulation parameter values in bold correspond to the reference values. 2SLS (two-stage least squares), DeLIVR, DeepIV, and Quantile IV (proposed method).
Figure 2
Figure 2
Association between common genetic variants at the SOST locus (chr17:41631099–42236156) and the exposure and outcomes considered in our MR study in the UK Biobank The variants are ordered with respect to their genomic coordinate, and the leftmost part of the plot shows the observed LD between the variants. The color lines represent association p values for every genetic variant and phenotype colored with respect to the sign of the regression coefficient (red for trait increasing and blue for trait decreasing). The sclerostin-decreasing allele is the coded allele throughout.
Figure 3
Figure 3
Quantile IV estimate of the average effect varying the levels of circulating sclerostin about the mean on heel bone mineral density and osteoporosis in the UK Biobank The shaded region corresponds to 90% bootstrap confidence intervals. The plots cover the central 99% of the exposure range.
Figure 4
Figure 4
Mendelian randomization estimates of a 1 SD decrease in circulating sclerostin levels on osteoporosis and cardiovascular diseases accounting for direct effects by rs113533733 in the UK Biobank
See this image and copyright information in PMC

References

    1. Lawlor D.A., Tilling K., Davey Smith G. Triangulation in Aetiological Epidemiology. Int. J. Epidemiol. 2016;45:1866–1886. doi: 10.1093/ije/dyw314. - DOI - PMC - PubMed
    1. Voight B.F., Peloso G.M., Orho-Melander M., Frikke-Schmidt R., Barbalic M., Jensen M.K., Hindy G., Hólm H., Ding E.L., Johnson T., et al. Plasma HDL Cholesterol and Risk of Myocardial Infarction: A Mendelian Randomisation Study. Lancet. 2012;380:572–580. doi: 10.1016/S0140-6736(12)60312-2. - DOI - PMC - PubMed
    1. Burgess S., Thompson S.G. Multivariable Mendelian Randomization: The Use of Pleiotropic Genetic Variants to Estimate Causal Effects. Am. J. Epidemiol. 2015;181:251–260. doi: 10.1093/aje/kwu283. - DOI - PMC - PubMed
    1. Triozzi J.L., Hsi R.S., Wang G., Akwo E.A., Wheless L., Chen H.-C., Tao R., Ikizler T.A., Robinson-Cohen C., Hung A.M., VA Million Veteran Program Mendelian Randomization Analysis of Genetic Proxies of Thiazide Diuretics and the Reduction of Kidney Stone Risk. JAMA Netw. Open. 2023;6 doi: 10.1001/jamanetworkopen.2023.43290. - DOI - PMC - PubMed
    1. Ference B.A., Robinson J.G., Brook R.D., Catapano A.L., Chapman M.J., Neff D.R., Voros S., Giugliano R.P., Davey Smith G., Fazio S., Sabatine M.S. Variation in PCSK9 and HMGCR and Risk of Cardiovascular Disease and Diabetes. N. Engl. J. Med. 2016;375:2144–2153. doi: 10.1056/NEJMoa1604304. - DOI - PubMed

MeSH terms