Sparse dimensionality reduction approaches in Mendelian randomisation with highly correlated exposures

被引:4
|
作者
Karageorgiou, Vasileios [1 ,2 ]
Gill, Dipender [1 ,3 ,4 ]
Bowden, Jack [2 ,4 ]
Zuber, Verena [1 ]
机构
[1] Imperial Coll London, Fac Med, Sch Publ Hlth, Dept Epidemiol & Biostat, London, England
[2] Univ Exeter, Exeter, England
[3] St Georges Univ London, Inst Infect & Immun, Dept Clin Pharmacol & Therapeut, London, England
[4] Novo Nordisk Res Ctr Oxford, Genet Dept, Oxford, England
来源
ELIFE | 2023年 / 12卷
关键词
Mendelian randomisation; principal component analysis; causal inference; coronary heart disease; Multivariable Mendelian Randomization; Dimensionality Reduction; Sparsity; PRINCIPAL COMPONENT ANALYSIS; VARIANTS;
D O I
10.7554/eLife.80063
中图分类号
Q [生物科学];
学科分类号
07 ; 0710 ; 09 ;
摘要
Multivariable Mendelian randomisation (MVMR) is an instrumental variable technique that generalises the MR framework for multiple exposures. Framed as a regression problem, it is subject to the pitfall of multicollinearity. The bias and efficiency of MVMR estimates thus depends heavily on the correlation of exposures. Dimensionality reduction techniques such as principal component analysis (PCA) provide transformations of all the included variables that are effectively uncorrelated. We propose the use of sparse PCA (sPCA) algorithms that create principal components of subsets of the exposures with the aim of providing more interpretable and reliable MR estimates. The approach consists of three steps. We first apply a sparse dimension reduction method and transform the variant-exposure summary statistics to principal components. We then choose a subset of the principal components based on data-driven cutoffs, and estimate their strength as instruments with an adjusted F-statistic. Finally, we perform MR with these transformed exposures. This pipeline is demonstrated in a simulation study of highly correlated exposures and an applied example using summary data from a genome-wide association study of 97 highly correlated lipid metabolites. As a positive control, we tested the causal associations of the transformed exposures on coronary heart disease (CHD). Compared to the conventional inverse-variance weighted MVMR method and a weak instrument robust MVMR method (MR GRAPPLE), sparse component analysis achieved a superior balance of sparsity and biologically insightful grouping of the lipid traits.
引用
收藏
页数:35
相关论文
共 50 条
  • [1] Mendelian randomisation approaches to the study of prenatal exposures: A systematic review
    Diemer, Elizabeth W.
    Labrecque, Jeremy A.
    Neumann, Alexander
    Tiemeier, Henning
    Swanson, Sonja A.
    PAEDIATRIC AND PERINATAL EPIDEMIOLOGY, 2021, 35 (01) : 130 - 142
  • [2] Mendelian randomisation with coarsened exposures
    Tudball, Matthew J.
    Bowden, Jack
    Hughes, Rachael A.
    Ly, Amanda
    Munafo, Marcus R.
    Tilling, Kate
    Zhao, Qingyuan
    Davey Smith, George
    GENETIC EPIDEMIOLOGY, 2021, 45 (03) : 338 - 350
  • [3] A Mendelian randomisation method for complex disease exposures
    Tudball, Matthew J.
    Bowden, Jack
    Smith, George Davey
    Tilling, Kate
    GENETIC EPIDEMIOLOGY, 2020, 44 (05) : 522 - 522
  • [4] Fused Lasso Dimensionality Reduction of Highly Correlated NWP Features
    Catalina, Alejandro
    Alaiz, Carlos M.
    Dorronsoro, Jose R.
    DATA ANALYTICS FOR RENEWABLE ENERGY INTEGRATION: TECHNOLOGIES, SYSTEMS AND SOCIETY (DARE 2018), 2018, 11325 : 13 - 26
  • [5] A novel multivariable Mendelian randomization framework to disentangle highly correlated exposures with application to metabolomics
    Chan, Lap Sum
    Malakhov, Mykhaylo M.
    Pan, Wei
    AMERICAN JOURNAL OF HUMAN GENETICS, 2024, 111 (09)
  • [6] Overlapping-sample Mendelian randomisation with multiple exposures: a Bayesian approach
    Linyi Zou
    Hui Guo
    Carlo Berzuini
    BMC Medical Research Methodology, 20
  • [7] Overlapping-sample Mendelian randomisation with multiple exposures: a Bayesian approach
    Zou, Linyi
    Guo, Hui
    Berzuini, Carlo
    BMC MEDICAL RESEARCH METHODOLOGY, 2020, 20 (01)
  • [8] Differences in causality between exposures and stroke subtypes: a mendelian randomisation approach
    Fill, J.
    Fokina, A.
    Klappacher, G.
    EUROPEAN HEART JOURNAL, 2023, 44
  • [9] Sparse Dimensionality Reduction Revisited
    Høgsgaard, Mikael Møller
    Kamma, Lior
    Larsen, Kasper Green
    Nelson, Jelani
    Schwiegelshohn, Chris
    arXiv, 2023,
  • [10] Comparison of Mendelian Randomisation and Bayesian Network Approaches for Causal Inference
    Howey, Richard
    Cordell, Heather J.
    HUMAN HEREDITY, 2017, 83 (05) : 235 - 235