Cohort Analytics Core

National Human Genome Research Institute

Investigators

Linked publications, trials & patents

Abstract

In its inaugural year, the Cohort Analytics Core has begun work on 20 collaboration projects, as well as developed various coding packages and pipelines to make genomic and phenomic analyses more accessible. We have also held training sessions for groups and mentored trainees throughout NHGRI to help get scientists acquainted with EHR-based analyses. Some highlights from our recent work include the following: Sickle cell trait (SCT) This project is a collaboration with Dr. Vence Bonhams group in the Social and Behavior Research Branch at NHGRI. Although individuals with sickle cell trait (SCT) harbor only a single HbS allele and typically remain asymptomatic, some studies have shown that SCT can be associated with adverse clinical manifestations. It is important to better understand these clinical outcomes, especially in making policies to protect student athletes who identify as Black/African American from both cardiovascular events and discrimination. Compared with those without HbS of predominantly African ancestry, our recent PheWAS in All of Us found that HbS heterozygotes were associated with increased risk for aseptic necrosis of bone and pulmonary embolism. However, after those with concomitant -hemoglobinopathies, such as HbC and -thalassemia alleles, were excluded, both associations decreased. The PheWAS also showed that SCT was associated with anemia during pregnancy, iron deficiency anemias , gout , and abnormal results of kidney function study. From this study, we suggest that it is important for clinicians to investigate concomitant -hemoglobinopathies to determine the risks of vascular occlusive complications in SCT. An abstract from this is accepted for ASHG 2023 and a manuscript is in progress. Uric Acid This project is a collaboration with Oleg Shchelochkov, M.D., Director of Medical Genetics and Genomics Medicine Training Program. Although hyperuricemia is observed to be associated with adverse cardiovascular outcomes, the causal relationship is unclear due to environmental confounding such as diet and lifestyle. With All of Us data, we found that although hypertensive heart diseases and congestive heart failure were strongly associated with normalized plasma uric acid levels directly derived from EHR, they were not associated with genetic risk score (GRS)-estimated uric acid levels calculated from GWAS-confirmed variants . As a positive control phenotype, gout was strongly associated with both EHR-derived and GRS-estimated uric acid levels. Dr. Shchelochkov will present these findings at the 2023 American Society of Human Genetics (ASHG) Annual Meeting. TTR Transthyretin amyloidosis (ATTR) is an adult-onset, multi-system disorder with diverse symptomatology determined by the localization of amyloid plaques. The p.V122I/V142I variant of TTR is particularly frequent in African Americans, with an estimated allele frequency of 3-4%. Previous research has suggested an association of this variant with incident heart failure among black individuals. Our PheWAS of p.V142I across all genotyped individuals in All of Us found nominal associations with peripheral nerve disorders, retinal drusen, and disorders of the kidney and ureters consistent with known phenotypic presentations of ATTR. Our PheWAS of p.V142I for individuals over the age of 60 with predominantly African ancestry similarly showed only a variety of nominally significant phenotype associations. Neither PheWAS showed significant associations with heart failure phenotypes. Preliminary survival analysis data for heart failure individuals based on their genotype at this variant position also showed no significant difference, as did survival analysis for heart failure individuals with peripheral neuropathy. We are currently developing a PheRS for ATTR using curated characteristics taken from the OMIM clinical synopsis. CLYBL/CD320 and Vitamin B12 Vitamin B12 is an essential cofactor in humans that we obtain from our diet. Severe B12 deficiency is associated with anemia and neurological impairments, while mild B12 deficiency is very common. Studies have found that a gene called CLYBL plays a vital role in the Vitamin B12 metabolism. Its loss is found in 2.7 % of all human chromosomes due to a single nucleotide polymorphism (SNP rs41281112) that changes Arginine (Arg259) to a stop codon and leads to loss of the protein. The homozygosity of this mutation is associated with a strong reduction of circulating Vitamin B12 levels. We conducted PheWAS on a variant of this gene that is believed to be pathogenic. To create phecodes, we mapped the electronic health record (EHR) data using the ICD to phecodes map. Using the whole genome sequencing data (WGS) across ancestries in the NIHs All of Us Research Program, our logistic regression models found phecodes such as hypotony of eye, macroglobulinemia, megaloblastic anemia, Vitamin B-complex deficiencies, circadian rhythm sleep disorders and cystic fibrosis to be positively associated in homozygous individuals for CLYBL. Further studies are in progress to explore our findings such as analyzing laboratory data in All of Us on these individuals and conducting ancestry specific PheWAS studies. TP53 Polyadenylation signal In collaboration with the Biesecker lab, a PheWAS was conducted on the TP53 polyadenylation signal-disrupting variant chr17:7668434:T:G (Hg38) to elucidate its associations with distinct phenotypes. We performed variant extraction from 245,378 participants within the All of Us database and identified 3383 individuals who were heterozygous and <20 who were homozygous for the G-allele. As a quality control, we determined that the heterozygous carriers have a mean variant allele fraction (VAF) 0.4988 +/- 0.0792, indicating that most of the variants are likely germline. Our preliminary analyses demonstrate that the polyadenylation signal-disrupting variant of TP53 is consistently associated with higher risk of neoplasms and tissue overgrowth in multiple ancestry groups. With the diverse background of All of Us data, our findings not only validate the previously reported associations from GWAS studies, but also argue against linkage disequilibrium confounding, due to the consistency of the associations across ancestry groups. Autoinflammatory This is a collaboration with Dr. Dan Kastners group and led by summer intern student Saee Risbud. Autoinflammatory disease are caused by changes in the bodys innate immune system. The P/LP variants in ADA2, MEFV, MVK, PSTPIP1, TNFRSF1A, and UBA1 were observed to be associated with various autoinflammatory diseases. The PheWAS on ADA p.Arg169Gln variant with 155 heterozygous carriers did not show significant phenotype association, which is consistent with the recessive nature of this variant. The PheWAS on MEFV p.Val726Ala variant showed nominal association with keratitis and uveitis. The PheWAS on NOD2 truncated variants showed significant association with Crohns disease. This study demonstrated the utilization of genotype-first phenotyping in autoinflammatory diseases. Further collaboration with Dr. Kastners group has been planned. RUNX1 and DDX41 This project is a collaboration of Dr. Calvo from the Dept. Laboratory Medicine at Clinical Center. Inherited pathogenic/likely pathogenic (P/LP) variants of RUNX1 lead to familial platelet disorder with predisposition to myeloid malignancy (FPDMM), and those of DDX41 lead to late onset myeloid neoplasms. However, previous generated data has mostly focused on clinically ascertained probands and their families and thus may be subjected to ascertainment biases. Here, we used a genotype-first strategy to investigate the natural behaviors of the P/LP variants involving RUNX1 and DDX41 in a phenotypically unselected population on the All of Us Research Program. This