Bone Density study with SNP Genotypes (imputed) data on dbGaP (3,355 participants)

phs002102 — Bone Microarchitecture

<p>Osteoporosis affects more than 28 million people in the United States and the lifetime risk for osteoporosis-related morbidity is higher than a woman's combined risk for breast cancer, endometrial cancer and ovarian cancer. Health care expenditures for osteoporotic patients in this country are currently nearly 13 billion dollars per annum and are predicted to increase markedly in the next decades due the aging of the population; therefore, it is important to understand the factors that contribute to bone strength and fracture risk.  With the advent of skeletal imaging modalities such as high resolution peripheral quantitative computed tomography (HR-pQCT), it is now possible to study the genetics of more highly refined peripheral skeletal microarchitecture phenotypes. Because these refined phenotypes have not been measured in many cohort studies, this project was a collaborative effort to include almost all the existing data on HR-pQCT of the radius and tibia, and genetics from around the world.  Cohorts involved in the discovery included:  1) <a href="https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000007">Framingham Osteoporosis Study</a>, 2) Mayo Clinic cohort; 3) Geneva Retirees cohort; 4) OFELY cohort from Lyon France; 5) STRAMBO cohort from Lyon France; 6) Swedish Male cohorts.  Replication cohorts undergoing de novo genotyping include:  1) QUALYOR cohort from Lyon France; 2) CaMOS cohort from Canada.  Other than the Framingham cohort and the Swedish male cohorts, all discovery cohorts underwent genome wide genotyping with the Affymetrix Axiom Biobank array that had common variants along with exome content based on the early findings from the Exome Sequencing Project.  All genotype data from the discovery cohorts were then imputed using the Haplotype Reference Consortium reference panel.  Variants for replication genotyping using the Kasp technology were selected based on novelty compared with previously identified loci using DXA phenotypes, minor allele frequency, underlying LD structure within a locus, and likelihood of a variant being functional as assessed using various algorithms for coding and non-coding variants.    </p>