Here you find references to all published mGWAS in humans, ordered by publication date. If a study is missing from this list, please let me know. Please note that I do not apply a strict inclusion criterion for what I would call a GWAS with metabolomics, or rather a GWAS with biochemical traits (such as a GWAS with cholesterol levels). The latter are well covered by the GWAS catalogue.
This table was initially published in Kastenmüller et al., Genetics of human metabolism: an update. Hum. Mol. Genet. 2015 and has been updated as of 13 February 2025. If you are interested in associations at specific loci you may also use the SNiPA block annotation tool [read this post], and if you are looking for associations specific to certain metabolites, try the Metabolomics GWAS Server [read this post].
| Biofluid | Metabolic traits | Platform | Study population | # Traits | Cohort size | # Loci | Reference |
| Serum | Targeted MS | MS (Biocrates) | German | 363 + all ratios | 284 | 4 | Gieger et al., 2008 |
| Plasma and Serum | Sphingolipids | MS | European | 33 + 43 ratios | 4,400 | 5 | Hicks et al., 2009 |
| Serum | Mainly phospholipids | MS (Biocrates) | German, British | 163 + 26,406 ratios | 1,809 + 422 | 9 | Illig et al., 2010 |
| Urine | NMR-derived metabolites | NMR (Chenomx) | German | 59 + 1,661 ratios | 862 + 992 | 5 | Suhre et al., 2011 |
| Plasma | n-3 PUFAs (ALA, EPA, DPA, and DHA) | Gas chromatography | European ancestry (CHARGE) | 4 | 8,866 | 3 | Lemaitre et al., 2011 |
| Serum | Non-targeted MS, knowns | MS (Metabolon) | German, British | 276 + 37 179 ratios | 1,768 + 1,052 | 37 | Suhre et al., 2011 |
| Urine and Plasma | Urine: NMR peaks, plasma: mainly phospholipids | NMR + MS | British | Urine: 512 peaks plasma: 163 + ratios | 211 | 3 | Nicholson et al., 2011 |
| Serum | Mainly lipid traits | NMR (Brainshake) | Finnish | 117 + 99 ratios | 8,330 | 31 | Kettunen et al., 2012 |
| Plasma | Phospholipids + sphingolipids | MS (Biocrates) | European | 153 | 4,034 | 35 | Demirkan et al., 2012 |
| Serum | Mainly lipid traits | NMR | Finnish | 117 + 99 ratios | 8,330 | 30 | Tukiainen et al., 2012 |
| Serum | Mainly lipid traits and low-weight metabolites | NMR (Brainshake) | Finnish, British | 130 | 1,905 + 4,703 | 34 | Inouye et al., 2012 |
| Serum | Non-targeted MS, unknowns | MS (Metabolon) | German | 517 | 1,768 | 34 | Krumsiek et al., 2012 |
| Urine | NMR peaks | NMR | Brazil | 2425 | 265 | 2 | Montoliu et al., 2013 |
| Plasma | NMR peaks | NMR | German | 8,600 + 124,750 ratios | 1,757 | 7 | Raffler et al., 2013 |
| Plasma | Fatty acids 16:0, 16:1n-7, 18:0, and 18:1n-9 | GC | 5 European cohorts | 4 | 8,961 | 7 | Wu et al., 2013 |
| Serum | MS peaks | MS | Swedish | 6,138 | 402 + 489 | 7 | Hong et al., 2013 |
| Plasma | Amino acids, amines, polar metabolites, lipids | MS | USA (European ancestry) | 217 | 2,076 | 31 | Rhee et al., 2013 |
| Urine | NMR peaks | NMR | European, Brazil | 1,276 | 835 + 601 | 11 | Rueedi et al., 2014 |
| Serum | Non-targeted MS, knowns and unknowns | MS | African American | 308 | 1,260 | 19 | Yu et al., 2014 |
| Serum | Non-targeted MS, knowns and unknowns | MS (Metabolon) | European | 486 + 98,346 ratios | 7,824 | 145 | Shin et al., 2014 |
| Serum | Targeted MS (mainly phospholipids) + non-targeted MS (knowns) | MS (Metabolon & Biocrates) | European | 344 (151 + 193) | 1,809 + 843 | 12 new | Ried et al., 2014 |
| Serum | NMR-derived metabolites | NMR | European | 42 | 2,118 | 8 | Demirkan et al., 2015 |
| Serum | Mainly phospholipids | MS (Biocrates) | European | 129 | 7,478 + 1,182 | 31 | Draisma et al., 2015 |
| Blood | Circulating trans fatty acids | Diverse | 7 cohorts of European-ancestry, replication in 3 non-European cohorts | >5-7 (study dependent) | 8,013 + (1,082 + 669 + 657) | 1 (31 SNPs in FADS1/2 cluster) | Mozaffarian et al., 2015 |
| Urine | NMR-derived metabolites and NMR peaks | NMR | European | 15,379 features (incl. ratios) | 3,861 + 1,691 | 26 | Raffler et al., 2015 |
| Blood | Amino Acids and Acylcarnitines | MS | European | 96 | 2,107 | 16 | Burkhardt et al., 2015 |
| Plasma | FIA-MS derived, mainly acylcarnitines and amino acids | MS | US population | 63 | 1,490 + 2,022 | 6 | Kraus et al., 2015 |
| RBC | Fatty acids in RBCs | GC | Framingham Heart Study | 14 | 2,633 | 5 | Tintle et al., 2015 |
| Blood | Lipoprotein lipids and subclasses, fatty acids, amino acids, glycolysis precursors | NMR (Brainshake) | 14 cohorts from Europe | 123 | 24,925 | 62 | Kettunen et al., 2016 |
| Plasma | Non-targeted MS | LC-MS | European-American | 217 | 2,076 + 1,528 | 4 | Rhee et al., 2016 |
| Plasma | Non-targeted MS focusing on carnitine pathways with whole genome sequencing | LC-MS | European-American | 16 | 1,456 | 2 | Yazdani et al., 2016 |
| Plasma | Non-targeted MS with exome sequencing | LC-MS (Metabolon) | African-Americans | 308 | 1,361 + 508 | 4 | Yu et al., 2016 |
| Plasma | Caffeine metabolites | UPLC- MS/MS or UPLC- ESI-MS/MS | six population-based studies of European ancestry | 4 + 1 ratio | 9,876 | 4 | Cornelius et al., 2016 |
| Plasma + RBCs | 16:1n-7, 18:1n-7, 18:1n-9, 20:1n-9, 22:1n-9, 24:1n-9 fatty acids | gas chromatography or gas-liquid chromatography | Meta-analysis of Chinese and European ancestries | 6 | 3,521 + 12,020 | 10 | Hu et al., 2017 |
| Plasma | Small peptide subset of the non-targeted Metabolon metabolomics platform | Metabolon HD4 | European and African Americans (ARIC study) | 25 | 1,552 + 1,872 | 22 | de Vries et al., 2017 |
| Plasma | Targeted and Non-targeted MS | Biocrates + Metabolon | TwinsUK study | 160 + 488 | 1,001 twins | 7 | Yet et al., 2016 |
| Plasma | Non-targeted MS | Metabolon HD4 | TwinsUK study | 644 | 1,960 with three time points | 101 (17 genes with potential LoF variants) | Long et al., 2017 |
| Plasma | Non-targeted MS | Metabolon HD4 | Qatari | 826 | 614 (exome seq) + 382 (2.5 Omni array) | 21 common + 12 rare variants | Yousri et al., 2018 |
| Plasma | Amino acids | NMR | non-diabetic Finnish men | 9 | 8,545 | 5 common + 2 rare/gene-based | Teslovich et al., 2018 |
| Serum | Non-targeted MS | Lipid-related metabolites from Metabolon platform | ARIC | 102 | 1552 European-and 1872 African-Americans, using WES and WGS | 7 novel lipid-related metabolite-genomic region associations | Feofanova et al., Genetics, 2018 |
| Plasma | Amino acids, acylcarnitines, sphingolipids | MS | Singapore Chinese | 162 + 70 ratios | 2,434 | 17 | Chai et al., ASHG 2018 abstract |
| Plasma | Lipidomics | Lipotype platform | European | 141 + 12 lipid classes | 2,045 | 35 lipidspecies- associated loci | Tabassum et al., Nature Comms, 2019, preprint BioRxiv, 2018 (presented at ASHG 2018) |
| Plasma + serum | Glycine | Metabolon HD4 + Biocrates p180 + Nightingale | European | 1 | 80,003 | 27 | Wittemans et al., 2019 |
| Serum | Non-targeted MS | Metabolon HD4 | Middle-aged individuals (45 to 64 years) from the ARIC study | 122 | 1,376 | 52 Loss of function variants (Exome sequencing, p<1E-6) | Yazdani et al., 2019 |
| NA | NA | NA | LIFE-Adult & LIFE-Heart studies | 8 steroid hormones | 7,667 | 15 novel | Pott et al., J Clin Endocrinol Metab. 2019 |
| Plasma | Plasma triglycerides, phospholipids | LC-MS | Leiden Longevity Study, Netherlands Twin Register, Erasmus Rucphen Family (ERF) study | 90 lipids | 5,537 | 34 replicated + 11 novel | Demirkan et al., bioRxiv 2019 |
| Serum | Non-targeted MS | Metabolon HD4 | Elite athletes | 751 | 490 | 19 elite athlete mQTLs, 4 endurance mQTLs | Al-Khelaifi et al., Scientific Reports, 2019 |
| Urine | Non-targeted MS | Metabolon HD2/HD4 | Patients with reduced kidney function | 1,172 | 1,627 | 240 | Schlosser et al., Nature Genetics, 2020 |
| Saliva | Non-targeted MS | Metabolon HD4 | TwinsUK + SHIP | 476 | 1,419 | 14 metabolites at 11 loci | Abhishek et al., Hum Mol Gen, 2020 |
| Serum | Non-targeted MS | Metabolon HD4 | two cohorts (Israeli) | 1,170 | 413 | 68 | Bar et al., Nature, 2020 |
| Serum | PUFA, omega-3, omega-6, and LA and DHA | NMR | PREDIMED Plus trial, Valencia, Spain | 5 | 426 | 1 genome-wide sign. locus (FADS1 region) | Coltell et al., Nutrients 2020 |
| Serum | Non-targeted MS | Metabolon HD4 | Hispanic (Hispanic Community Health Study/Study of Latinos participants) | 640 | 3,926 | 46 | Feofanova et al., AJHG 2020 |
| Cerebrospinal fluid (CSF) | Non-targeted MS | Metabolon HD4 | WADRC and WRAP Alzheimer's studies | 338 | 689 | 16 | Panyard et al., Nature Comms, 2021, previously bioRxiv 2020 |
| Plasma | Non-targeted MS | in-house LC-MS/MS platform for polar lipophilic metabolites | FINRISK02 + replication in Framingham Heart Study | 11,067 features | 7,013 + 2,886 | 118 loci associated with 2,319 features | Qin et al., medRxiv 2020 |
| Serum | Direct Infusion High-Resolution Mass Spectrometry (DIHRMS) | in-house platform described here | Pakistan Risk of Myocardial Infarction Study (PROMIS) + British blood donors from the INTERVAL study | 340 lipid metabolites in PROMIS + 399 in INTERVAL | 5,662 in PROMIS + 13,814 in INTERVAL | 253 lipidQTLs at 24 independent loci in PROMIS + 502 lipid QTLs at 38 loci in INTERVAL | Harshfield et al., BMC Medicine 2021, preprint on medRxiv 2020 |
| ? | ? | ? | TOPMed | ? | ? | Multi-Ethnic Whole-Genome Sequencing Analysis of Human Metabolome Identifies 93 Novel Genetic Loci | Bing Yu et al., ASHG 2020 |
| Blood | Multi-platform | Biocrates p180, Nightingale NMR, Metabolon HD4 | Fenland, EPIC-Norfolk, INTERVAL | 174 | 86,507 | 499 associations at 144 loci | Lotta et al., Nature Genetics, 2021, preprint bioRxiv, 2020 |
| Plasma | N-acyl ethanolamine and ceramide plasma lipidome | Targeted Lipidomics using UPLC/ESI-MS/MS | UK residents, families | 9 N-acyl ethanolamine and 16 ceramide species | 999 members of 196 British Caucasian families | ~5 | McGurk et al., Human Mol Gen, 2021 |
| Plasma | Non-targeted MS | not specified | Han Chinese | 161 + 12,280 ratios | 1,028 + 523 | 13 | Wang et al., Clinical and Translational Medicine, 2021 |
| Urine | Non-targeted MS | Metabolon HD4 | German Chronic Kidney Disease (GCKD) cohort | 1487 + 53,714 ratios | 4,864 (4,795 for ratios) | 420 associations (MAF<1%), 38 gene associations for rare variant tests (60 for ratios) | Cheng et al., Nature Communications, 2021 |
| Blood | Targeted & non-targeted MS (study also analyzed transciptomics & proteomics) | Biocrates p150 & Metabolon HD2 | DIRECT study | 163 (Biocrates) + 223 (Metabolon) | 3,029 | 301 independent metabolite-QTLs | Vinuela et al., medRxiv 2021 |
| Serum | Non-targeted MS | Metabolon HD4 | African American Study of Kidney Disease and Hypertension | 652 | 619 | 42 | Luo et al., Kidney Int., 2021 |
| Urine | 1H NMR | in-house | Caucasian | 1,276 features after QC | 555 | 21 | Flitman et al., J. Proteome Res., 2021, preprint bioRxiv 2020 |
| Plasma | UPLC–QTOFMS | West Coast Metabolomics Center at University of California Davis | Old Order Amish founder population | 355 lipid species | 650 | five rare Amish enriched loci, three of which are novel, replicated 7 previously known common loci | Montasser et al., bioRxiv 2021 |
| Plasma, mitochondrial DNA | Targeted MS | Biocrates p150 | KORA study | 151 | 2,718 | NA (mitochondrial DNA) | Aboulmaouahib et al. (Human Molecular Genetics, 2021) |
| Plasma | Non-targeted MS | Metabolon HD4 | NIHR BioResource (only individuals of full European ancestry) | 722 named metabolites | 8,809 | 74 novel | Hysi et al. (Metabolites, 2022) |
| Plasma, before/after liquid meal | Lipoprotein lipids and subclasses, fatty acids, amino acids, glycolysis precursors | NMR (Nightingale) | NEO study, Leiden, Netherlands | 148 | up to 4,734 | several QTLs with fasting-to-postprandial-state changes | Li-Gao et al. (Diabetes, 2021) |
| Plasma | Non-targeted MS | Broad institute | Chronic Renal Insufficiency Cohort (CRIC) Study (822 White and 687 Black study participants) | 537 | 1678 | 45 mQTL at 19 loci | Rhee et al. (Kidney Intl, 2022) |
| Plasma | Non-targeted MS | Metabolon HD4 | Men from the late-settlement region of Finland, bottle-necked population | 1,391 | 6,136 | 303 novel association signals | Yin et al. (Nat. Comms 2022), medRxiv 2021 |
| Serum | LC-MS | in-house | Korea Association REsource (KARE) project | 64 (53 with pathway information) | 627 | 32 (17 with glycine, 15 with dimethylglycine) | Jung et al. (Front. Genet., 2022) |
| Plasma | NMR | Nightingale platform (limited to 16 selected metabolites) | UK Biobank | 16 | 94,464 | 213 | Smith et al., bioRxiv 2022 |
| Plasma | Non-targeted MS | Metabolon HD4 | Interval study | 995 | 2,805 WGS + 3,924 WES | 40 novel associations implicating rare coding variants | Bomba et al., AJHG 2022 |
| Plasma | Non-targeted MS | Metabolon HD4, limited to acetaminophen metabolites | Qatar Biobank | 9 acetaminophen metabolites and 36 ratios between them | 520 | 2 | Thareja et al., Metabolites, 2022 |
| Plasma | Lipoprotein lipids and subclasses, fatty acids, amino acids, glycolysis precursors | NMR (Nightingale) | UK Biobank (white British UKB sub-population) | 168 traits + >14,000 ratios | 98,698 | 215 (out of 892 tested lipid risk loci, this is not a GWAS) | Suhre et al., medRxiv 2022 |
| Serum | Targeted MS with Whole Exome Sequencing | Biocrates p180 | Alpine CHRIS Cohort | 175 | 5505 | 85 | König et al., Metabolites 2022 |
| Plasma | fourteen PUFA and MUFA phenotypes | NMR (Nightingale) | UK Biobank + replication ion two external European studies | 14 | 101,729 | 51 | Francis et al., medRxiv 2022 |
| Blood | Lipoprotein lipids and subclasses, fatty acids, amino acids, glycolysis precursors | NMR (Nightingale) | 33 predominantly population-based cohorts | 233 | 136,016 | >400 | Karjalainen et al., medRxiv 2022 |
| Plasma | Non-targeted MS | Metabolon HD4 | Interval & EPIC Norfolk | 913 | 14,296 | 423 | Surendran et al., Nature Medicine 2022 |
| Plasma | Warfarin metabolites | HPLC-MS/MS | sub- Saharan black African population (South Africa, Uganda) | 13 | 548 | 373 unique SNPs in 13 genes | Asiimwe et al., Frontiers Pharmacology, 2022 |
| Plasma | Targeted MS | GC-MS | Japanese | 121 | 4,888 | 60 | Iwasaki et al., iScience 2023 |
| Plasma | Non-targeted MS | Metabolon HD4 | Canadian Longitudinal Study of Aging (CLSA) | 1,091 + 309 ratios | 8,299 | 1,509 | Chen et al., Nature Genetics 2023 |
| Plasma | Non-targeted MS | Uses two LC-MS methods on an Agilent system & whole genome sequencing | Black individuals from the Jackson Heart Study | 2,291 metabolite peaks | 2,466 | 519 | Tahir et al., Nature Comms 2022 |
| Red blood cells (RBC) | UHPLC-MS | in-house | Donors at REDS-III US blood centers | 382 | 243 | 423 | Moore et al., JBC 2022 |
| Skeletal muscle (duck) | Non-targeted MS | LC–MS/MS and SPME-GC-HRMS | Pekin duck × Liancheng duck crosses | 3431 metabolites and 702 volatiles | 423 | 2,862 signals | Liu et al., Advanced Science 2023 |
| Urine & Plasma | Non-targeted MS | Metabolon HD4 | German Chronic Kidney Disease (GCKD) study | 1,916 | 5,023 | 1,299 | Schlosser et al., Nat. Genet. 2023 |
| Urine | NMR | 600 MHz Bruker AVANCE III HD NMR spectrometer | 3 European cohorts (Generation Scotland, VIKING, FinnDiane) | 54 | 8,026 | 52 | Valo et al., medRxiv 2023 |
| Plasma | four MS/MS assays | amino acids, hormones, water-soluble vitamins (WSV), fat-soluble vitamins (FSV), and metal elements | Chinese pregnant women | 84 | 34,394 | 30 novel metabolite-gene associations | Liu et al., medRxiv 2023 |
| Plasma | Mass spectrometry-based shotgun lipidomics | Lipotype GmbH | Finnish individuals from the GeneRISK cohort | 179 lipids from 13 lipid classes | 7,174 | 495 genome-trait associations in 56 genetic loci | Ottensmann et al., Nat. Comm. 2023 |
| Plasma | LC-MS | plasma eicosanoids and related metabolites, including PUFAs | Atherosclerosis Risk in Communities (ARIC) Study | 223 | 8,406 | 41 | Rhee et al., Commun. Biol. 2023 |
| Cerebrospinal fluid (CSF) and brain | Non-targeted MS | Metabolon HD4 | multiple Alzheimer's cohorts | 440 | 2,602 (CSF) and 1,016 (brain) | 219 (CSF) and 36 (brain) | Wang et al., pre-print on Research Square, 2023 |
| Red blood cells (RBC) | Women's Health Initiative Memory study | 28 fatty acids | 7,479 women | 12 | Westra et al., PLEFA 2023 | ||
| Serum | NMR | Chenomx | Critically ill, mechanically ventilated patients with septic shock | 20 | 230 | 4 | Daubney et al., Crit. Care Expl. 2024 |
| Predominantly plasma | NMR | Nightingale | Meta-analysis of 33 cohorts with updated quantification version of Nightingale platform | 233 | 136,016 | >400 independent loci | Karjalainen et al., Nature 2024 |
| Plasma | NMR | Nightingale | Estonian Biobank and UK Biobank | 249 | 619,372 | 89,489 locus-metabolite pairs at 8,917 loci | Tambets et al., medRxiv 2024 |
| Plasma | NMR | Nightingale | UK Biobank | 249 | ~450,000 | 29,824 locus–metabolite associations mapping to 753 regions | Zoodsma et al., medRxiv 2025 |
| Urine | NMR | Nightingale Health urine NMR platform | 3 European cohorts | 54 | 8,011 | 26 loci (54 associations for 19 metabolites) | Valo et al., Nat. Comms 2025 |
| Blood | Non-targeted MS | Metabolon/Broad platform | TOPMed | up to 1,666 | up to 11,840 | 1985 novel, validated 761 known | Feofanova et al., Nat. Comms 2023 |
| Blood | Non-targeted MS | Metabolon HD4 | NEO and Rhineland Study | 224 with 10%-90% missingness | 594 + 4,165 | 55 associations with missingness | Faquih et al., medRxiv 2024 |
Tweet This
