Dataset
European ancestry samples in this dataset are included in the "DIAMANTE (European) T2D GWAS" dataset.
Publications
Genetic fine mapping and genomic annotation defines causal mechanisms at type 2 diabetes susceptibility loci.
Gaulton KJ, et al.
Nat Genet. 2015 Dec;47(12):1415-25. doi: 10.1038/ng.3437.
Low-frequency and rare exome chip variants associate with fasting glucose and type 2 diabetes susceptibility.
Wessel J, et al.
Nat Commun. 2015 Jan 29;6:5897. doi: 10.1038/ncomms6897.
Meta-analysis of genome-wide association studies in African Americans provides insights into the genetic architecture of type 2 diabetes.
Ng MC, et al.
PLoS Genet. 2014 Aug 7;10(8):e1004517. doi: 10.1371/journal.pgen.1004517.
Genetic background of patients from a university medical center in Manhattan: implications for personalized medicine.
Tayo BO, et al.
PLoS One. 2011 May 4;6(5):e19166. doi: 10.1371/journal.pone.0019166.
Dataset phenotypes
Phase 1 analysis:
- type 2 diabetes
- type 2 diabetes adjusted for BMI
- fasting glucose
- fasting glucose adjusted for BMI
- HbA1c
- HbA1c adjusted for BMI
Phase 2 analysis:
- serum creatinine
- systolic blood pressure
- HDL cholesterol
- eGFR-creat (serum creatinine)
- diastolic blood pressure
- LDL cholesterol
- BMI
Phase 1 analysis of diabetic complications:
- Chronic kidney disease in type 2 diabetes
- End-stage renal disease in type 2 diabetes
- Neuropathy in type 2 diabetes
Phase 2 analysis of diabetic complications:
- Coronary artery disease in type 2 diabetes
- Claudication in type 2 diabetes
- Hypertension in type 2 diabetes
- Myocardial infarction in type 2 diabetes
- Peripheral vascular disease in type 2 diabetes
- Stroke in type 2 diabetes
Dataset subjects
| Cases | Controls | Cohort (Click to view selection criteria for cases and controls) | Ancestry |
|---|---|---|---|
| 2,293 | 6,880 | The Charles Bronfman Institute for Personalized Medicine BioMe Biobank | Mixed |
Project
The Charles Bronfman Institute for Personalized Medicine (IPM) BioMe Biobank is a consented, EMR-linked medical care setting biorepository of the Mount Sinai Medical Center (MSMC) drawing from a population of over 70,000 inpatients and 800,000 outpatient visits annually. MSMC serves diverse local communities of upper Manhattan, including Central Harlem, East Harlem, and Upper East Side with broad health disparities.
The BioMe Biobank was founded in September 2007 and as of September 2016, > 34,000 participants were enrolled. IPM BioMe Biobank populations include 28% African American, 38% Hispanic Latino predominantly of Caribbean origin, and 23% Caucasian/White. Enrolled participants consent to be followed throughout their clinical care (past, present, and future) at Mount Sinai in real-time, integrating their genomic information with their electronic health records for discovery research and clinical care implementation. The BioMe Biobank disease burden is reflective of health disparities with broad public health impact. Biobank operations are fully integrated in clinical care processes, including direct recruitment from clinical sites waiting areas and phlebotomy stations by dedicated Biobank recruiters independent of clinical care providers, prior to or following a clinician standard of care visit. Recruitment currently occurs at a broad spectrum of over 30 clinical care sites.
Acknowledgments
The BioMe Biobank is supported by the Andrea and Charles Bronfman Philanthropies.
Experiment summary
The BioMe AMP T2D GWAS sample set is comprised of 13,034 unique individuals. The major ancestry groups are admixed American (41.5%) and African American (38%). Subjects of European ancestry comprise 19.8% of the set, and 0.006% are South Asian.
Samples were genotyped on at least one of three platforms: Illumina Exome Array, which assayed nearly 200,000 autosomal variants; Illumina GWAS Array, which assayed over 844,000 autosomal variants; and Affymetrix GWAS Array, which assayed over 837,000 autosomal variants.
Phenotypes were recorded during routine medical visits, and reflect the value at the time of the most recent visit at which that measurement was taken (or, where applicable, at the last non-pregnant visit at which that measurement was taken).
T2D case and control definition algorithms were developed by a multidisciplinary team of scientists, clinicians, and software specialists. Comprehensive documentation of the algorithms can be found at www.phekb.org/phenotype/type-2-diabetes-mellitus. This algorithm has been validated with excellent performance statistics; 100% sensitivity and > 98% positive predictive value for cases, and ≥ 98% sensitivity and ≥ 98% positive predictive value for controls.
In addition to T2D status, the phenotypes considered in the Phase 1 analysis were fasting glucose level, measured in blood samples taken from patients who had not had food or drink for at least 8 hours, and HbA1c level, which is an indicator of average blood glucose levels over a three-month period.
The Phase 2 analysis included seven traits: serum creatinine levels (4,844 samples), diastolic blood pressure (7,597 samples), systolic blood pressure (7,586 samples), LDL cholesterol levels (2,076 samples), HDL cholesterol levels (2,059 samples), body mass index (7,249 samples), and estimated glomerular filtration rate, calculated using serum creatinine levels (4,844 samples).
Phase 1 analysis of diabetic complications included the traits chronic kidney disease in type 2 diabetics (408 cases), end-stage renal disease in type 2 diabetics (214 cases), and diabetic neuropathy (501 cases).
Phase 2 analysis of diabetic complications included the traits Coronary artery disease in type 2 diabetics (1,063 cases), claudication in type 2 diabetics (559 cases), hypertension in type 2 diabetics (2,123 cases), myocardial infarction in type 2 diabetics (493 cases), peripheral vascular disease in type 2 diabetics (817 cases), and stroke in type 2 diabetics (400 cases).
Overview of analysis and results
Data were analyzed by the Analysis Team at the Accelerating Medicines Partnership Data Coordinating Center (AMP-DCC), Broad Institute, using Loamstream software and the AMP-DCC Data Analysis Pipeline. Results are summarized below; see the Quality Control and Analysis Reports (links below) for full details. After sample quality control (excluding samples flagged for non-type 2 diabetes, and removing duplicates where samples had been assayed on multiple platforms), results from the two GWAS arrays were combined in meta-analysis and results from the exome array were analyzed separately. A single set of p-values was produced using a MEGA analysis strategy, including all samples in a single association test.
T2D associations were calculated in two models, BMI-unadjusted and BMI-adjusted. In both models, rs7903146, within the well-known T2D risk gene TCF7L2, was the only association reaching genome-wide significance, with p-values of 6.22 × 10−10 and 4.81 × 10−10 in the BMI-unadjusted and BMI-adjusted models, respectively. This association had been identified previously, but no other associations among the top 20 in either model had been identified previously.
None of the fasting glucose or HbA1c associations in either the BMI-unadjusted or BMI-adjusted models reached genome-wide significance, and none of the top 20 associations for these phenotypes in either model had been identified previously.
No associations with serum creatinine levels reached genome-wide significance, and none of the top 20 associations had been previously identified.
Systolic blood pressure values were increased by 15 for individuals on hypertension medications. One association with systolic blood pressure reached genome-wide significance: rs1843784 in the region of the GALNT18 gene, with a p-value of 1.18 × 10−8. That association was not previously identified, nor were any of the other top 20 associations for systolic blood pressure.
One association with HDL cholesterol reached genome-wide significance: rs247616 in the region of the CETP gene, with a p-value of 1.33 × 10−13. That association was previously identified, as was one other of the top 20 associations for HDL cholesterol.
One association for estimated glomerular filtration rate (eGFR) reached genome-wide significance: rs717588 in the region of the PCNX2 gene, with a p-value of 2.36 × 10−8. That association was not previously identified, nor were any of the other top 20 associations for eGFR.
Diastolic blood pressure values were increased by 10 for individuals on hypertension medications. No associations with diastolic blood pressure reached genome-wide significance, and none of the top 20 associations had been previously identified.
For LDL cholesterol one association, rs7412, reached genome-wide significance in the well known APOE region on chromosome 19. One other LDL association among the top 20 had been previously identified.
No associations with BMI reached genome-wide significance, and none of the top 20 associations had been previously identified.
In the Phase 1 and Phase 2 analyses of diabetic complications, no significant associations were found (as expected, due to the small sample sizes for these traits).
Detailed reports
Genotype Data Quality Control Report (download PDF)
AMP-DCC Phase 1 Data Analysis Report (download PDF)
AMP-DCC Phase 2 Data Analysis Report (download PDF)
AMP-DCC Phase 1 Diabetic Complications Data Analysis Report (download PDF)
AMP-DCC Phase 2 Diabetic Complications Data Analysis Report (download PDF)
External Links to BioMe AMP T2D GWAS results
These data are also available in dbGaP under the following accessions: phs000388.v1.p1; phs000948.v1.p1; phs000925.v1.p1.