Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study
Abstract Background There is increasing need for effective incorporation of high-dimensional genetics data from individuals with varied ancestry in genome-wide association (GWAS) analyses. Classically, multi-ancestry GWAS analyses are performed using statistical meta-analysis to combine results cond...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-01-01
|
| Series: | BMC Genomics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12864-025-11229-1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832586025699901440 |
|---|---|
| author | Alan Kuang Marie-France Hivert M. Geoffrey Hayes William L. Lowe Denise M. Scholtens |
| author_facet | Alan Kuang Marie-France Hivert M. Geoffrey Hayes William L. Lowe Denise M. Scholtens |
| author_sort | Alan Kuang |
| collection | DOAJ |
| description | Abstract Background There is increasing need for effective incorporation of high-dimensional genetics data from individuals with varied ancestry in genome-wide association (GWAS) analyses. Classically, multi-ancestry GWAS analyses are performed using statistical meta-analysis to combine results conducted within homogeneous ancestry groups. The emergence of cosmopolitan reference panels makes collective preprocessing of GWAS data possible, but impact on downstream GWAS results in a mega-analysis framework merits investigation. We utilized GWAS data from the multi-national Hyperglycemia and Adverse Pregnancy Outcome Study to investigate differences in GWAS findings using a homogeneous ancestry meta-analysis versus a heterogeneous ancestry mega-analysis pipeline. Maternal fasting and 1-hr glucose and metabolomics measured during a 2-hr 75-gram oral glucose tolerance test during early third trimester pregnancy were evaluated as phenotypes. Results For the homogeneous ancestry meta-analysis pipeline, variant data were prepared by identifying sets of individuals with similar ancestry and imputing to ancestry-specific reference panels. GWAS was conducted within each ancestry group and results were combined using random-effects meta-analysis. For the heterogeneous ancestry mega-analysis pipeline, data for all individuals were collectively imputed to the Trans-Omics for Precision Medicine (TOPMed) cosmopolitan reference panel, and GWAS was conducted using a unified mega-analysis. The meta-analysis pipeline identified genome-wide significant associations for 15 variants in a region close to GCK on chromosome 7 with maternal fasting glucose and no significant findings for 1-hr glucose. Associations in this same region were identified using the mega-analysis pipeline, along with a well-documented association at MTNR1B on chromosome 11 with both fasting and 1-hr maternal glucose. For metabolomics analyses, the number of significant findings in the heterogeneous ancestry mega-analysis far exceeded those from the homogeneous ancestry meta-analysis and confirmed many previously documented associations, but genomic inflation factors were much more variable. Conclusions For multi-ancestry GWAS, heterogeneous ancestry mega-analysis generates a rich set of variants for analysis using a cosmopolitan reference panel and results in vastly more significant, biologically credible and previously documented associations than a homogeneous ancestry meta-analysis approach. Genomic inflation factors do indicate that findings from the mega-analysis pipeline may merit cautious interpretation and further follow-up. |
| format | Article |
| id | doaj-art-e83a191c059641f990cd9e04f44326db |
| institution | Kabale University |
| issn | 1471-2164 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Genomics |
| spelling | doaj-art-e83a191c059641f990cd9e04f44326db2025-01-26T12:16:35ZengBMCBMC Genomics1471-21642025-01-0126111510.1186/s12864-025-11229-1Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) studyAlan Kuang0Marie-France Hivert1M. Geoffrey Hayes2William L. Lowe3Denise M. Scholtens4Department of Preventive Medicine, Northwestern University Feinberg School of MedicineDepartment of Medicine, Massachusetts General HospitalDepartment of Medicine, Northwestern University Feinberg School of MedicineDepartment of Medicine, Northwestern University Feinberg School of MedicineDepartment of Preventive Medicine, Northwestern University Feinberg School of MedicineAbstract Background There is increasing need for effective incorporation of high-dimensional genetics data from individuals with varied ancestry in genome-wide association (GWAS) analyses. Classically, multi-ancestry GWAS analyses are performed using statistical meta-analysis to combine results conducted within homogeneous ancestry groups. The emergence of cosmopolitan reference panels makes collective preprocessing of GWAS data possible, but impact on downstream GWAS results in a mega-analysis framework merits investigation. We utilized GWAS data from the multi-national Hyperglycemia and Adverse Pregnancy Outcome Study to investigate differences in GWAS findings using a homogeneous ancestry meta-analysis versus a heterogeneous ancestry mega-analysis pipeline. Maternal fasting and 1-hr glucose and metabolomics measured during a 2-hr 75-gram oral glucose tolerance test during early third trimester pregnancy were evaluated as phenotypes. Results For the homogeneous ancestry meta-analysis pipeline, variant data were prepared by identifying sets of individuals with similar ancestry and imputing to ancestry-specific reference panels. GWAS was conducted within each ancestry group and results were combined using random-effects meta-analysis. For the heterogeneous ancestry mega-analysis pipeline, data for all individuals were collectively imputed to the Trans-Omics for Precision Medicine (TOPMed) cosmopolitan reference panel, and GWAS was conducted using a unified mega-analysis. The meta-analysis pipeline identified genome-wide significant associations for 15 variants in a region close to GCK on chromosome 7 with maternal fasting glucose and no significant findings for 1-hr glucose. Associations in this same region were identified using the mega-analysis pipeline, along with a well-documented association at MTNR1B on chromosome 11 with both fasting and 1-hr maternal glucose. For metabolomics analyses, the number of significant findings in the heterogeneous ancestry mega-analysis far exceeded those from the homogeneous ancestry meta-analysis and confirmed many previously documented associations, but genomic inflation factors were much more variable. Conclusions For multi-ancestry GWAS, heterogeneous ancestry mega-analysis generates a rich set of variants for analysis using a cosmopolitan reference panel and results in vastly more significant, biologically credible and previously documented associations than a homogeneous ancestry meta-analysis approach. Genomic inflation factors do indicate that findings from the mega-analysis pipeline may merit cautious interpretation and further follow-up.https://doi.org/10.1186/s12864-025-11229-1Genome-wide association analysisMulti-ancestryMeta-analysisMega-analysis |
| spellingShingle | Alan Kuang Marie-France Hivert M. Geoffrey Hayes William L. Lowe Denise M. Scholtens Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study BMC Genomics Genome-wide association analysis Multi-ancestry Meta-analysis Mega-analysis |
| title | Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study |
| title_full | Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study |
| title_fullStr | Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study |
| title_full_unstemmed | Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study |
| title_short | Multi-ancestry genome-wide association analyses: a comparison of meta- and mega-analyses in the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study |
| title_sort | multi ancestry genome wide association analyses a comparison of meta and mega analyses in the hyperglycemia and adverse pregnancy outcome hapo study |
| topic | Genome-wide association analysis Multi-ancestry Meta-analysis Mega-analysis |
| url | https://doi.org/10.1186/s12864-025-11229-1 |
| work_keys_str_mv | AT alankuang multiancestrygenomewideassociationanalysesacomparisonofmetaandmegaanalysesinthehyperglycemiaandadversepregnancyoutcomehapostudy AT mariefrancehivert multiancestrygenomewideassociationanalysesacomparisonofmetaandmegaanalysesinthehyperglycemiaandadversepregnancyoutcomehapostudy AT mgeoffreyhayes multiancestrygenomewideassociationanalysesacomparisonofmetaandmegaanalysesinthehyperglycemiaandadversepregnancyoutcomehapostudy AT williamllowe multiancestrygenomewideassociationanalysesacomparisonofmetaandmegaanalysesinthehyperglycemiaandadversepregnancyoutcomehapostudy AT denisemscholtens multiancestrygenomewideassociationanalysesacomparisonofmetaandmegaanalysesinthehyperglycemiaandadversepregnancyoutcomehapostudy |