Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

When the proteome meets the metabolome observational and Mendelian randomization analyses

Zheng, Rui ; Delgado-Velandia, Mario ; Ärnlöv, Johan ; Sundström, Johan ; Engström, Gunnar LU ; Smith, J. Gustav LU orcid ; Dekkers, Koen F. ; Lundmark, Per ; Fall, Tove and Lind, Lars (2026) In Metabolism: Clinical and Experimental 180.
Abstract

Objective The basis for protein synthesis is the genetic code. Many of these proteins will affect intermediary metabolites by acting as enzymes, hormones, or by other actions. The aim of the present study was to assess the relationships of a large number of proteins with endogenous metabolites. Methods Plasma protein levels were measured by the proximity extension assay (PEA) and metabolites by mass spectrometry. Cross-sectional relationships of 242 proteins and 790 metabolites were evaluated in the EpiHealth and POEM studies using a discovery/validation approach. Genetic instruments identified in UK Biobank for protein levels (n = 1621) and genetics for metabolite levels (n = 777) in SCAPIS and EpiHealth were employed for Mendelian... (More)

Objective The basis for protein synthesis is the genetic code. Many of these proteins will affect intermediary metabolites by acting as enzymes, hormones, or by other actions. The aim of the present study was to assess the relationships of a large number of proteins with endogenous metabolites. Methods Plasma protein levels were measured by the proximity extension assay (PEA) and metabolites by mass spectrometry. Cross-sectional relationships of 242 proteins and 790 metabolites were evaluated in the EpiHealth and POEM studies using a discovery/validation approach. Genetic instruments identified in UK Biobank for protein levels (n = 1621) and genetics for metabolite levels (n = 777) in SCAPIS and EpiHealth were employed for Mendelian randomization (MR) analysis regarding putative causal associations. Results In the observational analyses, 20% of the evaluated pairwise protein-metabolite associations were found significant in both the discovery and validation samples. We could however only find support for causal effects in the MR analysis for <0.1% of the pairwise associations, representing 326 unique proteins. The R2 for the relationship between the MR and observational estimates was only 0.05. 37 protein-metabolite relationships that were significant in a congruent fashion in both the observational and MR analyses were identified. A searchable online protein vs metabolite atlas was created for the scientific community to use these results. We also give some examples where metabolites were used to enhance protein findings in cardiovascular epidemiological research. Conclusion This study provides a comprehensive assessment of a large number of protein- metabolite relationships using both observational and MR analyses, highlighting how these results could be used to enhance clinical research.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Cardiovascular disease, Genetics, Metabolomics, Proteomics, UK Biobank
in
Metabolism: Clinical and Experimental
volume
180
article number
156602
publisher
Elsevier
external identifiers
  • scopus:105035103905
  • pmid:41962653
ISSN
0026-0495
DOI
10.1016/j.metabol.2026.156602
language
English
LU publication?
yes
id
7403c039-fee9-4cdd-acb6-315051fc9188
date added to LUP
2026-05-20 15:25:29
date last changed
2026-06-03 16:25:34
@article{7403c039-fee9-4cdd-acb6-315051fc9188,
  abstract     = {{<p>Objective The basis for protein synthesis is the genetic code. Many of these proteins will affect intermediary metabolites by acting as enzymes, hormones, or by other actions. The aim of the present study was to assess the relationships of a large number of proteins with endogenous metabolites. Methods Plasma protein levels were measured by the proximity extension assay (PEA) and metabolites by mass spectrometry. Cross-sectional relationships of 242 proteins and 790 metabolites were evaluated in the EpiHealth and POEM studies using a discovery/validation approach. Genetic instruments identified in UK Biobank for protein levels (n = 1621) and genetics for metabolite levels (n = 777) in SCAPIS and EpiHealth were employed for Mendelian randomization (MR) analysis regarding putative causal associations. Results In the observational analyses, 20% of the evaluated pairwise protein-metabolite associations were found significant in both the discovery and validation samples. We could however only find support for causal effects in the MR analysis for &lt;0.1% of the pairwise associations, representing 326 unique proteins. The R<sup>2</sup> for the relationship between the MR and observational estimates was only 0.05. 37 protein-metabolite relationships that were significant in a congruent fashion in both the observational and MR analyses were identified. A searchable online protein vs metabolite atlas was created for the scientific community to use these results. We also give some examples where metabolites were used to enhance protein findings in cardiovascular epidemiological research. Conclusion This study provides a comprehensive assessment of a large number of protein- metabolite relationships using both observational and MR analyses, highlighting how these results could be used to enhance clinical research.</p>}},
  author       = {{Zheng, Rui and Delgado-Velandia, Mario and Ärnlöv, Johan and Sundström, Johan and Engström, Gunnar and Smith, J. Gustav and Dekkers, Koen F. and Lundmark, Per and Fall, Tove and Lind, Lars}},
  issn         = {{0026-0495}},
  keywords     = {{Cardiovascular disease; Genetics; Metabolomics; Proteomics; UK Biobank}},
  language     = {{eng}},
  publisher    = {{Elsevier}},
  series       = {{Metabolism: Clinical and Experimental}},
  title        = {{When the proteome meets the metabolome observational and Mendelian randomization analyses}},
  url          = {{http://dx.doi.org/10.1016/j.metabol.2026.156602}},
  doi          = {{10.1016/j.metabol.2026.156602}},
  volume       = {{180}},
  year         = {{2026}},
}