Document text
Principal Investigator: Joshua Denny
Organization: NATIONAL HUMAN GENOME RESEARCH INSTITUTE
Fiscal Year: 2024
Award: $1,427,000
Funding agency: National Human Genome Research Institute
1. Blood pressure (BP) and hypertension (HTN) in large-scale biobanks
HTN is a major risk factor for cardiovascular disease, the leading cause of death worldwide. We conducted GWAS for blood pressure traits in over one million individuals, identifying 2103 associated loci, and leveraged our results to create a polygenic risk score that is a powerful predictor of HTN. The PRS predicted BP and HTN in individuals of predominantly of European and African genetic ancestry. This was published in Nature Genetics.
We also developed a generalizable method to evaluate drug-response with electronic health records (EHRs) in the All of Us Research Program (All of Us) and found differences in anti-HTN drug effectiveness by self-identified race and ethnicity. We discovered Black and Hispanic participants were started on anti-HTN medications at higher SBP than White participants. Seven out of 19 anti-HTN drugs were less effective in Black and six were less effective in Hispanic, compared to non-Hispanic White participants. This was published in Clinical Pharmacology & Therapeutics.
2. PheWAS tool development and General Biobank Data Evaluation
All of Us, with 287,012 participants sharing EHRs, was compared to UK Biobank (UKB) and the US general population to study disease prevalence. Among 314 diseases evaluated, 80.9% were more common in All of Us than in the US population, and 85.6% of 2,515 diseases were more prevalent in All of Us compared to UKB. Despite prevalence differences, All of Us proves valuable for studying a wide range of diseases. This was published in JAMIA.
We also developed PheTK, a python package that runs PheWAS analyses quickly and efficiently. Compared to the standard R PheWAS package, PheTK takes 1/3 less run time. PheTK can handle biobank level data, works with standard data formats (such as Hail matrix table and OMOP common data model), and supports both phecode 1.2 and phecodeX 1.0. PheTK has been released on GitHub and a paper describing it on medrxiv is in revision at Bioinformatics.
Another project uses phecodes and the Phenotype-Genotype Reference Map to study differences in phenotyping between patient provided survey data and EHR data within All of Us. We mapped survey questions to phecodes and compared performance to known genetic associations. This project aims to evaluate phenotyping performance to ultimately strengthen the statistical power and validity of biomedical and genetic research.
3. Additional medication studies in All of Us
Antiplatelet therapy with clopidogrel reduces subsequent events in patients with a recent myocardial infarction or percutaneous coronary procedure. Its efficacy is influenced by CYP2C19 variants. To study pharmacogenomic interactions in All of Us, we evaluated genetic associations with clopidogrel response in 1,937 participants. CYP2C19*2 was associated with a higher risk of recurrent cardiovascular events in trans-ancestry analysis.
Up to 70% of increased risk of end stage kidney disease (ESKD) in individuals of West African ancestry can be explained by risk variants (RV) APOL1. Adjusted Cox regression analysis showed lower ESKD risk in individuals with two APOL1 RV and SGLT2i medication exposure. Notably, no participants with two APOL1 RV and SGLT2is exposure developed ESKD during the follow-up time.
We also looked at adverse cardiac events associated with androgen deprivation therapy (ADT), which has been revolutionary in preventing reoccurrence of prostate cancer. Utilizing All of Us, we found that ADT increases major cardiac adverse event risk even after adjusting for cardiovascular risk factors. Additionally, we replicated a known association of ADT usage with QTc interval elongation.
4. Genome and Phenome-wide association studies (GWAS & PheWAS)
Uterine fibroids are common benign tumors with poorly understood etiology. Using a multi-ancestry GWAS meta-analyses, we identified 371 associated loci, 24 of which were novel. The predicted expression of 568 gene-tissue pairs had significant associations with fibroids, and of those, 131 genes were novel. Within uterine tissue analyses, we observed six significant novel gene associations: SULT1E1, TSGA10, SHMT1, RPS26, SLC25A17, and CD59.
We also leveraged All of Us data to investigate associations of predicted loss-of-function (pLoF) PTVs in BSN and APBA1 with body mass index across a population of diverse ancestry. PheWAS uncovered novel associations of BSN and APBA1 heterozygous pLoF carriers with various phenotypes. Specifically, BSN pLoF variants were associated with pulmonary HTN, atrial fibrillation, and anticoagulant use, while APBA1 pLoF variants were linked to disorders of the temporomandibular joint. These findings underscore the potential of large-scale biobanks in advancing genetic discovery.
In addition, we launched several genome-wide associations in primary hypothyroidism (PH), Graves disease, endometriosis, Guillain-Barré syndrome, and Ménière's disease. Importantly, these analyses include diverse populations. Many of these findings are being combined with other data sets in larger consortia.
5. Phenotype discovery via pLoF variants in the Genome
We identified all pLoF across the genome and performed gene-based PheWAS for each gene, starting with 260 genes with known disease associations caused by loss of function variants. This approach is being used to refine our methods and gauge the sensitivity of EHR information to detect different phenotypes with rare pLoF mutations. We replicated >70% of the known gene-phenotype associations among these genes. We then picked ~260 genes without known phenotype associations for potential discovery, and are exploring these with manual review, use of other biobanks, and other resources.
6. Infectious Disease Phenotyping
Infectious diseases are a common cause of mortality and morbidity but have not been explored using EHR data. In EHR phenotyping, integrating billing codes, test results and medications improves performance. This approach is rarely employed for seasonal respiratory infection research. We integrated these data types in All of Us to identify cohorts of participants with respiratory viral infections, including COVID-19, and validated these cohorts using CDC seasonal trends. We are now using these cohorts to support genetic susceptibility studies.
7. Rare Diseases
We identified 30,702 patients with documented diagnosis for 234 genetic diseases. For each, we mapped their clinical features to the OMOP concepts and calculated phenotype risk scores for each disease in all patients. We identified substantial disparities in the burden of diseases for many genetic diseases. For example, we observed a much higher burden of diseases of cystic fibrosis (CF) among patients self-reporting as Black/African Americans, compared to those self-reporting as White. The manuscript is in preparation.
We identified 5,013 carriers and 177,395 non-carriers for CF in the All of Us cohort. A total of 2,806 phenotypes were analyzed. No statistically significant associations were identified in carriers. Associations with CF-related conditions in carriers were several orders of magnitude lower than those found in homozygotes The manuscript is in preparation.
8. PheRS
Most rare variants are classified as variants of unknown significance (VUS), limiting their clinical utility. We used phenotype risk scores (PheRS) in the All of Us cohort to classify rare variants (allele frequency <1%) in 81 genes from the ACMG secondary findings list. By evaluating 2.67 million rare variants in 200,435 participants, we established criteria based on PheRS to identify variants with strong evidence of pathogenicity ("PheRS-path") or benignity ("PheRS-benign"). This study highlights the utility of PheRS and diverse cohorts in high-throughput variant classification.
Terms: <AIDP><APBA1><APBA1 gene><APOL-I><APOL1><APOL1 gene><Active Follow-up><Acute Autoimmune Neuropathy><Acute Infective Polyneuritis><Acute Inflammatory Demyelinating Polyradiculoneuropathy><Acute Inflammatory Polyneuropathy><Acute Inflammatory Polyradiculoneuropathy><African><African American group><African American individual><African American people><African American population><African Americans><African ancestry><African descent><Airway infections><All of Us Program><All of Us Research Program><All of Us Research Project><Allele Frequency><Amyloid Beta A4 Precursor Protein-Binding, Family A, Member 1><Anti-Hypertensive Agents><Anti-Hypertensive Drugs><Anti-Hypertensives><Anticoagulant Agents><Anticoagulant Drugs><Anticoagulants><AoURP><Atrial Fibrillation><Auricular Fibrillation><BMI><BMI percentile><BMI z-score><Basedow's Disease><Benign><Bio-Informatics><Bioinformatics><Biomedical Research><Black><Black race><Blood Platelets><Blood Pressure><Body Tissues><Body mass index><COVID-19><CV-19><CYP2C><CYP2C19><CYP2C19 gene><Cardiac><Cardiac infarction><Cardiovascular><Cardiovascular Body System><Cardiovascular Diseases><Cardiovascular Organ System><Cardiovascular system><Causality><Cause of Death><Classification><Clinical><Clinical Pharmacology><Clinics and Hospitals><Clinics or Hospitals><Code><Coding System><Communicable Diseases><Coronary><Coronavirus Infectious Disease 2019><Cystic Fibrosis><Cytochrome P450, Subfamily IIC, Polypeptide 19><D9S411E><Data><Data Set><Diagnosis><Disease><Disorder><Disparities><Disparity><Document Type><Drugs><ESRD><Effectiveness><Electronic Health Record><End stage renal failure><End-Stage Kidney Disease><End-Stage Renal Disease><Environment><Ethnic Origin><Ethnicity><Etiology><European><Evaluation><Event><Exophthalmic Goiter><Fibroid><Fibroid Neoplasm><Fibroid Tumor><Fibroid Uterus><Fibromyoma><GWA study><GWAS><Gene Frequency><Gene variant><General Population><General Public><Genes><Genetic><Genetic Diseases><Genetic Predisposition><Genetic Predisposition to Disease><Genetic Research><Genetic Susceptibility><Genetic analyses><Genetic propensity><Genome><Genotype><Goals><Graves' Disease><Guillain Barré Syndrome><Guillaine-Barre Syndrome><Heart Vascular><Heterozygote><Hispanic><Homozygote><Hypertension><Hypotensive Agent><Hypotensive Drugs><Hypothyroidism><Individual><Infectious Disease Pathway><Infectious Diseases><Infectious Disorder><Inherited Predisposition><Inherited Susceptibility><Investigation><Joint Diseases><LIN10><Laboratories><Landry's paralysis><Landry-Guillain-Barre Syndrome><Leiomyomatous Neoplasm><Leiomyomatous Tumor><Life Style><Lifestyle><Link><MINT1><MUNC18-1-Interaction Protein><Malignant Tumor of the Prostate><Malignant neoplasm of prostate><Malignant prostatic tumor><Manuals><Manuscripts><Maps><Marrow platelet><Medication><Mendelian randomization><Meniere's Disease><Meniere's Disorder><Meniere's Syndrome><Mephenytoin 4-Prime Hydroxylase><Meta-Analysis><Methods><Morbidity><Morbidity - disease rate><Mucoviscidosis><Myocardial Infarct><Myocardial Infarction><Ménière's disease><Nature><Non-Hispanic><Nonhispanic><Not Hispanic or Latino><Orphan Disease><P450C2C><Paper><Participant><Pathogenicity><Patient Self-Report><Patients><Performance><Pharmaceutical Preparations><Pharmacogenomics><Phenotype><Platelets><Population><Population Heterogeneity><Precision Health><Preparation><Prevalence><Procedures><Prostate CA><Prostate Cancer><Prostate malignancy><Prostatic Cancer><Provider><Publishing><Pulmonary Hypertension><Pythons><Quetelet index><Race><Races><Rare Diseases><Rare Disorder><Recurrence><Recurrent><Regression Analyses><Regression Analysis><Regression Diagnostics><Report (document)><Research><Research Resources><Resources><Respiratory Infections><Respiratory Tract Infections><Risk><Risk-associated variant><Running><Sample Size><Seasons><Self-Report><Source><Statistical Regression><Survey Instrument><Surveys><Systematics><TMJ Diseases><TMJ Disorders><TMJD><Temporomandibular Disorders><Temporomandibular Joint Diseases><Temporomandibular Joint Disorders><Temporomandibular Joint and Muscle Disorder><Test Result><Testing><Therapeutic><Thrombocytes><Time><Tissues><Uterine Body Fibroid><Uterine Body Leiomyoma><Uterine Corpus Fibroid><Uterine Corpus Leiomyoma><Uterine Fibroids><Uterine Fibroma><Uterine Leiomyoma><Uterus><Uterus Fibroma><Variant><Variation><Vascular Hypertensive Disease><Vascular Hypertensive Disorder><Vertebrate LIN10 Homolog><Viral Respiratory Tract Infection><Work><X11><X11-Alpha><X11-α><active followup><acute idiopathic polyneuritis><acute post-infectious polyneuropathy><acute postinfectious polyneuropathy><adverse event risk><allele variant><allelic frequency><allelic variant><ancestry analysis><androgen ablation therapy><androgen blockade therapy><androgen deprivation therapy><androgen deprivation treatment><anti-hypertension><arthropathic><arthropathies><arthropathy><biobank><biorepository><blood thinner><burden of disease><burden of illness><cardiac infarct><cardiovascular disorder><cardiovascular risk><cardiovascular risk factor><causation><circulatory system><clopidogrel><cohort><coronary attack><coronary infarct><coronary infarction><coronavirus disease 2019><coronavirus disease-19><coronavirus infectious disease-19><corpus uteri fibroid><corpus uteri leiomyoma><cost effective><data format><data integration><data modeling><data standardization><data standards><disease burden><disease causation><disease diagnosis><disease phenotype><disease risk><disorder risk><disparity in health><diverse populations><drug discovery><drug repositioning><drug repurposing><drug/agent><electronic health care record><electronic health medical record><electronic health plan record><electronic health registry><electronic medical health record><endometriosis><follow up><follow-up><followed up><followup><genetic analysis><genetic architecture><genetic association><genetic condition><genetic disorder><genetic etiology><genetic mechanism of disease><genetic variant><genetic vulnerability><genetically predisposed><genome wide association><genome wide association scan><genome wide association studies><genome wide association study><genomewide association scan><genomewide association studies><genomewide association study><genomic variant><health disparity><heart attack><heart infarct><heart infarction><heterogeneous population><heterozygosity><high blood pressure><high risk><hyperpiesia><hyperpiesis><hypertensive disease><hypertensive disorder><improved><individual heterogeneity><individual variability><individual variation><joint disorder><loss of function><loss of function mutation><model of data><model the data><modeling of the data><mortality><multi-modality><multimodality><novel><orphan disorder><personalization of treatment><personalized medicine><personalized therapy><personalized treatment><phenome><phenotypic data><polygenic risk score><population diversity><precision medicine><precision-based medicine><preparations><prevent><preventing><programs><racial><racial background><racial origin><rare allele><rare mutation><rare variant><repurposing agent><repurposing medication><response><risk allele><risk gene><risk genotype><risk loci><risk locus><risk variant><thrombopoiesis inhibitor><tool development><trait><translational study><trend><tumor><unclassified variant><uterus leiomyoma><variant of uncertain clinical significance><variant of uncertain significance><variant of undetermined significance><variant of unknown significance><viral respiratory infection><whole genome association analysis><whole genome association studies><whole genome association study><womb>