The number of genome-wide association studies (GWAS) with deep molecular phenotypes, especially metabolomics and proteomics, is rapidly increasing. On other places of this blog I maintain tables of published GWAS with metabolomics (mQTLs), proteomics (pQTLs), and DNA methylation (meQTLs).
Here I extend this collection by a table of all published epigenome-wide association studies (EWAS) with metabolomics, proteomics, and other deep molecular phenotypes. I include studies with a smaller set of traits (e.g. blood lipids) if they are of interest due to their large sample size and findings (i.e. reporting loci that overlap with multiomics traits).
Please read our review Connecting the epigenome, metabolome and proteome for a deeper understanding of disease (Suhre et al., J. Int. Medicine, 2021).
You may also be interested in the following resources:
Should you know of any study missing here, please let me know.
|Reference||Study Population/Cohort||Number of Samples||CpG-Trait Associations||Omics Phenotype|
|Petersen et al. (Hum. Mol. Genet. 2014)||German (KORA) population study||1814||20 (excluding CpGs with genetic confounders)||649 blood metabolic and lipid traits (Metabolon, Biocrates, and Numares/Lipofit platforms)|
|Sayols-Baixeras (Hum. Molec. Genet., 2016)||REGICOR study and Framingham Offspring Study (replication)||645+2542 (replication)||14 (9 genes + 2 intergenic loci)||Lipid traits (total, low-density and high-density lipoprotein cholesterol, and triglycerides)|
|Braun et al. (Clin. Epigenet., 2017)||Rotterdam Study||725+760 (replication)||5||blood levels of triglycerides, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and total cholesterol|
|Nano et al. (Gastroenterology, 2017)||Rotterdam Study||731+719 (replication)||8 (4 replicated)||blood for liver enzyme levels, including gamma-glutamyl transferase (GGT), alanine aminotransferase (ALT), and aspartate aminotransferase (AST)|
|Ahsan et al. (PLoS Genet., 2017)||Northern Sweden Population Health Study (NSPHS)||between 698 and 1033||124||121 protein biomarkers (Proseek immuno-assay technology)|
|Wahl et al. (BBA, 2018)||4 cohorts||3000+||7 (5 sites)||IgG glycosylation (24 glycans measured by UPLC, 50 by LC/MS)|
|Orozco et al.. (Hum. Mol. Gen., 2018)||Metabolic Syndrome in Men cohort||201||13 clinical traits at 21 loci||32 clinical traits in subcutaneous abdominal adipose tissue|
|Huang et al. (Epigenomics, 2018)||Emory Twin Study (ETS)||180 male twins||NA||12 annotated smoking-related metabolites identified from a metabolome-wide association study|
|Tindula et al. (Environmental Epigenetics, 2019)||Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) cohort||81 mother-child pairs||7 (after Bonferroni correction)||92 metabolites measured by targeted metabolomics (SRM monitoring)|
|Hillary et al. (Nat. Comm., 2019)||Lothian Birth Cohort 1936||750 healthy older adults||26 sites with the levels of 9 proteins||92 protein levels from the Olink neurology panel|
|Huan et al. (Nat.Comm. 2019)||Framingham Heart Study||4170||92 putatively|
causal CpGs for CVD traits by Mendelian randomization
|Zaghlool et al. (Nat. Comm., 2020)||German (KORA) population study and multi-ethnic QMDiab study||944 + 344 (replication)||98 (replicated)||1123 blood circulating proteins (Somalogic aptamers)|
|Hillary et al. (Genome Medicine 2020), pre-print on BioRxiv||Lothian Birth Cohort 1936||876||3||70 blood circulating protein levels (Olink inflammation panel)|
|Huan et al. (Epigenetics, 2020)||Framingham Heart Study||3565||227 CpGs with 40 nearby miRNAs (cis-miR-eQTMs)||283 blood circulating microRNAs (miRNA)|
|Battram et al., bioRxiv 2020||Avon Longitudinal Study of Parents and Children (ALSPAC)||940 (mothers)||no Bonferroni sign. hit; 29 associations at P<1E-7||400 “selected” ALSPAC traits (out of 2408), incl. Nightingale NMR data and body fat composition|
|Gomez‑Alonso et al. (Clinical Epigenetics, 2021)||KORA, LOLIPOP, NFBC1966, and YFS cohorts||5414||161 associations, covering 16 CpG sites at 11 loci and 57 metabolic measures||226 mostly lipid-related metabolic measures (NMR based)|
|Goodrich et al. (Epigenetics Insights 2021)||Mother-child cohorts from 3 maternity hospitals in Mexico City (ELEMENT study)||238 children (ages 8-14 years)||76 with LINE-1 DNA methylation, 27 with HSD11B2, 103 with H19, and 4 with IGF2||3758 serum metabolite features (574 of which are
|Yao et al. (Clinical Epigenetics, 2021)||Framingham Heart Study, replication in KORA+InCHIANTI+BLSA||4170 + 783 + 500 + 150 (replication)||10% of 16,416 cis CpG-transcript pairs from|
the discovery sample replicated
|Gene expression, using Illumina & Affymetrix platforms|
|Jhun et al. (Nat. Comms., 2021)||Meta-analysis of 15 EWASin Europeans, African Americans, and Hispanics||16265||148, 35, and 4 novel associations among Europeans, African Americans, and Hispanics, respectively||Blood lipids (HDL, LDL, TG)|
|Nie et al. (J. Hazardous Materials, 2021)||Non-smokers of the 2nd follow-up of the Wuhan-Zhuhai cohort||212||only CpGs associated with fasting glucose were reported (?)||10 Polycyclic aromatic hydrocarbon (PAH) metabolites|
(updated 11 July 2021)