Abstract
Multivariable Mendelian randomization allows simultaneous estimation of direct causal effects of multiple exposure variables on an outcome. When the exposure variables of interest are quantitative omic features, obtaining complete data can be economically and technically challenging: the measurement cost is high, and the measurement devices may have inherent detection limits. In this paper, we propose a valid and efficient method to handle unmeasured and undetectable values of the exposure variables in a one-sample multivariable Mendelian randomization analysis with individual-level data. We estimate the direct causal effects with maximum likelihood estimation and develop an expectation-maximization algorithm to compute the estimators. We show the advantages of the proposed method through simulation studies and provide an application to the Hispanic Community Health Study/Study of Latinos, which has a large amount of unmeasured exposure data.
| Original language | English (US) |
|---|---|
| Article number | 100338 |
| Journal | Human Genetics and Genomics Advances |
| Volume | 5 |
| Issue number | 4 |
| DOIs | |
| State | Published - Oct 10 2024 |
Funding
The authors are grateful for the support from the National Heart, Lung, and Blood Institute (NHLBI) under R01HL143885 and the National Human Genome Research Institute under R01HG009974 to develop this methodology. The authors are also grateful for funding from the National Institute on Aging under P30AG066615 and the Population Architecture using Genomics and Epidemiology Study under R01HG010297. The authors thank the staff and participants of HCHS/SOL for their important contributions (investigators\u2019 website: https://sites.cscc.unc.edu/hchs/). The HCHS/SOL is a collaborative study supported by contracts from the NHLBI to the University of North Carolina (HHSN268201300001I/N01-HC-65233), University of Miami (HHSN268201300004I/N01-HC-65234), Albert Einstein College of Medicine (HHSN268201300002I/N01-HC-65235), University of Illinois at Chicago (HHSN268201300003I/N01-HC-65236 Northwestern University), and San Diego State University (HHSN268201300005I/N01-HC-65237). The Genetic Analysis Center at the University of Washington was supported by NHLBI and NIDCR contracts (HHSN268201300005C AM03 and MOD03). Support for HCHS/SOL metabolomics data was graciously provided by the JLH Foundation (Houston, Texas). The authors also thank the Trans-Omics for Precision Medicine (TOPMed) program imputation panel (v.TOPMed-r2), which is supported by the NHLBI (see www.nhlbiwgs.org). The panel was constructed and implemented by the TOPMed Informatics Research Center at the University of Michigan (3R01HL-117626-02S1; contract HHSN268201800002I). The TOPMed Data Coordinating Center (3R01HL-120393-02S1; contract HHSN268201800001I) provided additional data management, sample identity checks, and overall program coordination and support. We gratefully acknowledge the participants who provided biological samples and the studies that provided data for TOPMed. D.Y.L. conceived the proposed method. Y.L. derived the mathematical formulas and developed the statistical programs to implement the method. Y.L. K.Y.W. and D.Y.L. designed and performed the simulation studies and verified the results. Y.L. A.G.H. P.G.L. H.M.H. M.G. K.E.N. C.G.D. C.L.A. B.Y. K.L.Y. V.L.B. R.K. L.H. B.T.J. Q.Q. T.S. and J.Y.M. acquired, processed, and analyzed the data and discussed the results. Y.L. K.Y.W. and D.Y.L. wrote the manuscript. All authors reviewed and commented on the manuscript. The authors declare no competing interests. The authors are grateful for the support from the National Heart, Lung, and Blood Institute (NHLBI) under R01HL143885 and the National Human Genome Research Institute under R01HG009974 to develop this methodology. The authors are also grateful for funding from the National Institute on Aging under P30AG066615 and the Population Architecture using Genomics and Epidemiology Study under R01HG010297 . The authors thank the staff and participants of HCHS/SOL for their important contributions (investigators\u2019 website: https://sites.cscc.unc.edu/hchs/ ). The HCHS/SOL is a collaborative study supported by contracts from the NHLBI to the University of North Carolina ( HHSN268201300001I/N01-HC-65233 ), University of Miami ( HHSN268201300004I/N01-HC-65234 ), Albert Einstein College of Medicine ( HHSN268201300002I/N01-HC-65235 ), University of Illinois at Chicago ( HHSN268201300003I/N01-HC-65236 Northwestern University), and San Diego State University ( HHSN268201300005I/N01-HC-65237 ). The Genetic Analysis Center at the University of Washington was supported by NHLBI and NIDCR contracts ( HHSN268201300005C AM03 and MOD03 ). Support for HCHS/SOL metabolomics data was graciously provided by the JLH Foundation (Houston, Texas). The authors also thank the Trans-Omics for Precision Medicine (TOPMed) program imputation panel (v.TOPMed-r2), which is supported by the NHLBI (see www.nhlbiwgs.org ). The panel was constructed and implemented by the TOPMed Informatics Research Center at the University of Michigan (3R01HL-117626-02S1; contract HHSN268201800002I). The TOPMed Data Coordinating Center (3R01HL-120393-02S1; contract HHSN268201800001I) provided additional data management, sample identity checks, and overall program coordination and support. We gratefully acknowledge the participants who provided biological samples and the studies that provided data for TOPMed.
Keywords
- causal inference
- correlated exposures
- detection limits
- instrumental variables
- missing data
- unmeasured confounders
ASJC Scopus subject areas
- Molecular Medicine
- Genetics(clinical)