Skip to navigation Skip to main content Skip to footer

Approved research

Prediction models of health and wellness traits based on genetics, anthropometric, and clinical data

Principal Investigator: Dr Nicole Washington
Approved Research ID: 40436
Approval date: July 24th 2018

Lay summary

Helix is a personal genomics software platform, formed by Illumina and two venture capital firms in 2015, with the mission to empower every person to improve their life through DNA. Helix sequences our users' Exome+, a combination of clinical-grade exome sequencing plus sequences that overlap SNPs traditionally found in genotyping arrays. Helix enables every person to access interpretations from their germline genomes through third party partners such as National Geographic, Mayo Clinic, Mount Sinai Sema4, with products that cover a broad spectrum of traits and diseases including ancestry, nutrition and diet, familial hypercholesterolemia, carrier screening, among others. Genome-wide association study (GWAS) is a standard method for identifying genetic risk factors for common traits and complex disease that hastens hypothesis generation and exploration of their biological underpinnings. However, individual variant-phenotype associations have limited personal utility. Genetic risk scores, a single predictive score for the overall genetic contribution to a trait that can be derived from GWAS analyses, are a powerful tool to stratify disease-risk populations for common traits and disease, including height, sleep, coronary-artery disease, and obesity. However, the catalog of these scores is limited and must be expanded and refined to maximize their utility. Helix's research goals include improving analytical methods and interpretability of the genome, furthering the overall scientific knowledge and genetic mechanisms of health and disease while growing the global knowledge of genetic-phenotype associations. Therefore, we aim to (1) develop and validate risk scores for common health and wellness-related traits and (2) develop novel genomic prediction algorithms using advanced machine learning techniques for linking genotype to phenotype. The UK Biobank (UKB) data is ideal for developing genetic risk scores and other prediction algorithms because of the large sample size and breadth of phenotypic data. Validation of association analysis from an independent cohort will strengthen findings and provide confidence in predictive models. With our population health study collaborations, such as with the Integrated Health Institute (IHI) and Renown Health, we will be able to perform validation studies in an effort to improve health outcomes. Together, the UK Biobank and Helix datasets will provide a rich analytical basis with which to discover and validate predictive models of health and wellness traits. We anticipate that this research effort will, in turn, enable more products to return actionable and individualized results, which we hope will increase the popularity, diversity, and impact of personal genomics.