Genomic data in the All of Us Research Program

Comprehensively mapping the genetic basis of human disease across diverse individuals is a long-standing goal for the field of human genetics1-4. The All of Us Research Program is a longitudinal cohort study aiming to enrol a diverse group of at least one million individuals across the USA to accelerate biomedical research and improve human health5,6. Here we describe the programme's genomics data release of 245,388 clinical-grade genome sequences. This resource is unique in its diversity as 77% of participants are from communities that are historically under-represented in biomedical research and 46% are individuals from under-represented racial and ethnic minorities. All of Us identified more than 1 billion genetic variants, including more than 275 million previously unreported genetic variants, more than 3.9 million of which had coding consequences. Leveraging linkage between genomic data and the longitudinal electronic health record, we evaluated 3,724 genetic variants associated with 117 diseases and found high replication rates across both participants of European ancestry and participants of African ancestry. Summary-level data are publicly available, and individual-level data can be accessed by researchers through the All of Us Researcher Workbench using a unique data passport model with a median time from initial researcher registration to data access of 29 hours. We anticipate that this diverse dataset will advance the promise of genomic medicine for all.

The frequency of pathogenic variation in the All of Us cohort reveals ancestry-driven disparities

Disparities in data underlying clinical genomic interpretation is an acknowledged problem, but there is a paucity of data demonstrating it. The All of Us Research Program is collecting data including whole-genome sequences, health records, and surveys for at least a million participants with diverse ancestry and access to healthcare, representing one of the largest biomedical research repositories of its kind. Here, we examine pathogenic and likely pathogenic variants that were identified in the All of Us cohort. The European ancestry subgroup showed the highest overall rate of pathogenic variation, with 2.26% of participants having a pathogenic variant. Other ancestry groups had lower rates of pathogenic variation, including 1.62% for the African ancestry group and 1.32% in the Latino/Admixed American ancestry group. Pathogenic variants were most frequently observed in genes related to Breast/Ovarian Cancer or Hypercholesterolemia. Variant frequencies in many genes were consistent with the data from the public gnomAD database, with some notable exceptions resolved using gnomAD subsets. Differences in pathogenic variant frequency observed between ancestral groups generally indicate biases of ascertainment of knowledge about those variants, but some deviations may be indicative of differences in disease prevalence. This work will allow targeted precision medicine efforts at revealed disparities.

Selection, optimization and validation of ten chronic disease polygenic risk scores for clinical implementation in diverse US populations

Polygenic risk scores (PRSs) have improved in predictive performance, but several challenges remain to be addressed before PRSs can be implemented in the clinic, including reduced predictive performance of PRSs in diverse populations, and the interpretation and communication of genetic results to both providers and patients. To address these challenges, the National Human Genome Research Institute-funded Electronic Medical Records and Genomics (eMERGE) Network has developed a framework and pipeline for return of a PRS-based genome-informed risk assessment to 25,000 diverse adults and children as part of a clinical study. From an initial list of 23 conditions, ten were selected for implementation based on PRS performance, medical actionability and potential clinical utility, including cardiometabolic diseases and cancer. Standardized metrics were considered in the selection process, with additional consideration given to strength of evidence in African and Hispanic populations. We then developed a pipeline for clinical PRS implementation (score transfer to a clinical laboratory, validation and verification of score performance), and used genetic ancestry to calibrate PRS mean and variance, utilizing genetically diverse data from 13,475 participants of the All of Us Research Program cohort to train and test model parameters. Finally, we created a framework for regulatory compliance and developed a PRS clinical report for return to providers and for inclusion in an additional genome-informed risk assessment. The initial experience from eMERGE can inform the approach needed to implement PRS-based testing in diverse clinical settings.

Demonstrating paths for unlocking the value of cloud genomics through cross cohort analysis

Recently, large scale genomic projects such as All of Us and the UK Biobank have introduced a new research paradigm where data are stored centrally in cloud-based Trusted Research Environments (TREs). To characterize the advantages and drawbacks of different TRE attributes in facilitating cross-cohort analysis, we conduct a Genome-Wide Association Study of standard lipid measures using two approaches: meta-analysis and pooled analysis. Comparison of full summary data from both approaches with an external study shows strong correlation of known loci with lipid levels (R2 ~ 83-97%). Importantly, 90 variants meet the significance threshold only in the meta-analysis and 64 variants are significant only in pooled analysis, with approximately 20% of variants in each of those groups being most prevalent in non-European, non-Asian ancestry individuals. These findings have important implications, as technical and policy choices lead to cross-cohort analyses generating similar, but not identical results, particularly for non-European ancestral populations.

Whole-genome sequencing as an investigational device for return of hereditary disease risk and pharmacogenomic results as part of the All of Us Research Program

Background

The All of Us Research Program (AoURP, “the program”) is an initiative, sponsored by the National Institutes of Health (NIH), that aims to enroll one million people (or more) across the USA. Through repeated engagement of participants, a research resource is being created to enable a variety of future observational and interventional studies. The program has also committed to genomic data generation and returning important health-related information to participants.

Methods

Whole-genome sequencing (WGS), variant calling processes, data interpretation, and return-of-results procedures had to be created and receive an Investigational Device Exemption (IDE) from the United States Food and Drug Administration (FDA). The performance of the entire workflow was assessed through the largest known cross-center, WGS-based, validation activity that was refined iteratively through interactions with the FDA over many months.

Results

The accuracy and precision of the WGS process as a device for the return of certain health-related genomic results was determined to be sufficient, and an IDE was granted.

Conclusions

We present here both the process of navigating the IDE application process with the FDA and the results of the validation study as a guide to future projects which may need to follow a similar path. Changes to the program in the future will be covered in supplementary submissions to the IDE and will support additional variant classes, sample types, and any expansion to the reportable regions.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13073-022-01031-z.

The Finland-United States investigation of non-insulin-dependent diabetes mellitus genetics (FUSION) study. I. An autosomal genome scan for genes that predispose to type 2 diabetes

We performed a genome scan at an average resolution of 8 cM in 719 Finnish sib pairs with type 2 diabetes. Our strongest results are for chromosome 20, where we observe a weighted maximum LOD score (MLS) of 2.15 at map position 69.5 cM from pter and secondary weighted LOD-score peaks of 2.04 at 56.5 cM and 1.99 at 17.5 cM. Our next largest MLS is for chromosome 11 (MLS = 1.75 at 84.0 cM), followed by chromosomes 2 (MLS = 0.87 at 5.5 cM), 10 (MLS = 0.77 at 75.0 cM), and 6 (MLS = 0.61 at 112.5 cM), all under an additive model. When we condition on chromosome 2 at 8.5 cM, the MLS for chromosome 20 increases to 5.50 at 69.0 cM (P=.0014). An ordered-subsets analysis based on families with high or low diabetes-related quantitative traits yielded results that support the possible existence of disease-predisposing genes on chromosomes 6 and 10. Genomewide linkage-disequilibrium analysis using microsatellite marker data revealed strong evidence of association for D22S423 (P=.00007). Further analyses are being carried out to confirm and to refine the location of these putative diabetes-predisposing genes.

The Finland-United States investigation of non-insulin-dependent diabetes mellitus genetics (FUSION) study. II. An autosomal genome scan for diabetes-related quantitative-trait loci

Type 2 diabetes mellitus is a complex disorder encompassing multiple metabolic defects. We report results from an autosomal genome scan for type 2 diabetes-related quantitative traits in 580 Finnish families ascertained for an affected sibling pair and analyzed by the variance components-based quantitative-trait locus (QTL) linkage approach. We analyzed diabetic and nondiabetic subjects separately, because of the possible impact of disease on the traits of interest. In diabetic individuals, our strongest results were observed on chromosomes 3 (fasting C-peptide/glucose: maximum LOD score [MLS] = 3.13 at 53.0 cM) and 13 (body-mass index: MLS = 3.28 at 5.0 cM). In nondiabetic individuals, the strongest results were observed on chromosomes 10 (acute insulin response: MLS = 3.11 at 21.0 cM), 13 (2-h insulin: MLS = 2.86 at 65.5 cM), and 17 (fasting insulin/glucose ratio: MLS = 3.20 at 9.0 cM). In several cases, there was evidence for overlapping signals between diabetic and nondiabetic individuals; therefore we performed joint analyses. In these joint analyses, we observed strong signals for chromosomes 3 (body-mass index: MLS = 3.43 at 59.5 cM), 17 (empirical insulin-resistance index: MLS = 3.61 at 0.0 cM), and 19 (empirical insulin-resistance index: MLS = 2.80 at 74.5 cM). Integrating genome-scan results from the companion article by Ghosh et al., we identify several regions that may harbor susceptibility genes for type 2 diabetes in the Finnish population.

Measurement of the planum temporale (PT) on magnetic resonance imaging scans: temporal PT alone and with parietal extension

The planum temporale (PT) has been of interest because of (1) its consistent left greater than right asymmetry among right-handed and most left-handed normal individuals; and (2) its relation to language, another variable shown to be highly left-lateralized in normal subjects. Individuals with neurodevelopmental disorders have been reported to show abnormal PT asymmetry (either reversed or absent asymmetry). Several studies have been conducted measuring the PT on MRI scans, although the results do not always concur. We review some of these studies and discuss methodological differences between them. Additionally, we propose a method that has proved to be highly reliable for the measurement of both temporal PT and its parietal extension (PT+).

Methionine synthase: high-resolution mapping of the human gene and evaluation as a candidate locus for neural tube defects

Periconceptual folate supplementation has been found to prevent the occurrence of many neural tube defects (NTDs). Consequently, genetic variation in folate metabolism genes is expected to contribute to the risk for neural tube defects. Methionine synthase catalyzes the vitamin B(12)-dependent conversion of homocysteine and 5-methyltetrahydrofolate to methionine and tetrahydrofolate. The observation that homocysteine and vitamin B(12) levels are independent predictors of NTD risk suggested that methionine synthase could be a candidate gene for NTDs. To assess the role of the MS gene in NTDs, we performed high-resolution physical mapping of the MS locus, isolated highly polymorphic markers linked to the MS gene, and tested for an association between specific MS alleles and NTDs. We mapped the MS gene to a position between 909 and 913 cR(10000) on chromosome 1 by radiation hybrid mapping. Polymorphic markers D1S1567 and D1S1568 map to locations no more than 900 and 194 kb from the MS gene, respectively. The segregation of these polymorphic markers was measured in 85 Irish NTD families. No allele of either marker showed a significant association with NTDs using the transmission disequilibrium test. A lack of association was also observed for the D1919G missense mutation within the gene. Our results suggest that inherited variation in the MS gene does not contribute to NTD risk in this population.

Type 2 diabetes: evidence for linkage on chromosome 20 in 716 Finnish affected sib pairs

We are conducting a genome scan at an average resolution of 10 centimorgans (cM) for type 2 diabetes susceptibility genes in 716 affected sib pairs from 477 Finnish families. To date, our best evidence for linkage is on chromosome 20 with potentially separable peaks located on both the long and short arms. The unweighted multipoint maximum logarithm of odds score (MLS) was 3.08 on 20p (location, chi = 19.5 cM) under an additive model, whereas the weighted MLS was 2.06 on 20q (chi = 57 cM, recurrence risk,lambda(s) = 1. 25, P = 0.009). Weighted logarithm of odds scores of 2.00 (chi = 69.5 cM, P = 0.010) and 1.92 (chi = 18.5 cM, P = 0.013) were also observed. Ordered subset analyses based on sibships with extreme mean values of diabetes-related quantitative traits yielded sets of families who contributed disproportionately to the peaks. Two-hour glucose levels in offspring of diabetic individuals gave a MLS of 2. 12 (P = 0.0018) at 9.5 cM. Evidence from this and other studies suggests at least two diabetes-susceptibility genes on chromosome 20. We have also screened the gene for maturity-onset diabetes of the young 1, hepatic nuclear factor 4-a (HNF-4alpha) in 64 affected sibships with evidence for high chromosomal sharing at its location on chromosome 20q. We found no evidence that sequence changes in this gene accounted for the linkage results we observed.

A large sample of finnish diabetic sib-pairs reveals no evidence for a non-insulin-dependent diabetes mellitus susceptibility locus at 2qter

In the first reported positive result from a genome scan for non-insulin-dependent diabetes mellitus (NIDDM), Hanis et al. found significant evidence of linkage for NIDDM on chromosome 2q37 and named the putative disease locus NIDDM1 (Hanis et al. 1996. Nat. Genet. 13:161-166). Their total sample was comprised of 440 Mexican-American affected sib-pairs from 246 sibships. The strongest evidence for linkage was at marker D2S125 and best estimates of lambdas (risk to siblings of probands/population prevalence) using this marker were 1.37 under an additive model and 1.36 under a multiplicative model. We examined this chromosomal region using linkage analysis in a Finnish sample comprised of 709 affected sib-pairs from 472 sibships. We excluded this region in our sample (multipoint logarithm of odds score </= -2) for lambdas >/= 1.37. We discuss possible reasons why linkage to 2q37 was not found and conclude that this region is unlikely to be playing a major role in NIDDM susceptibility in the Finnish Caucasian population.

Methods for precise sizing, automated binning of alleles, and reduction of error rates in large-scale genotyping using fluorescently labeled dinucleotide markers. FUSION (Finland-U.S. Investigation of NIDDM Genetics) Study Group

Large-scale genotyping is required to generate dense identity-by-descent maps to map genes for human complex disease. In some studies the number of genotypes needed can approach or even exceed 1 million. Generally, linkage and linkage disequilibrium analyses depend on clear allele identification and subsequent allele frequency estimation. Accurate grouping or categorization of each allele in the sample (allele calling or binning) is therefore an absolute requirement. Hence, a genotyping system that can reliably achieve this is necessary. In the case of affected sib-pair analysis without parents, the need for accurate allele calling is even more critical. We describe methods that permit precise sizing of alleles across multiple gels using the fluorescence-based, Applied Biosystems (ABI) genotyping technology and discuss ways to reduce genotyping error rates. Using database utilities, we show how to minimize intergel allele size variation, to combine data effectively from different models of ABI sequencing machines, and automatically bin alleles. The final data can then be converted into a format ready for analysis by statistical genetic packages such as MENDEL.