NEW YORK (GenomeWeb) – A new genome-wide association study has highlighted hundreds of new and known loci with ties to low bone mineral density (BMD), a well-recognized risk factor for osteoporosis and associated bone breaks, leading to a proposed polygenic risk score for osteoporosis.
"Knowledge of one's genetic risk for low BMD could be combined with clinical risk factors to identify patients, athletes, or military personnel at high risk for osteoporosis and fracture," wrote the study's sole author Stuart Kim, a developmental biology researcher at Stanford University.
For the analysis, published online today in PLOS One, Kim performed a GWAS based on genotyping data for nearly 400,000 UK Biobank participants with available BMD estimates. The search led to independent bone mineral density-associated SNPs at nearly 900 loci. From there, he did a series of analyses to train a BMD prediction algorithm, developing a genetic predictor for BMD and osteoporosis that was complemented by individuals' height, weight, sex, and age data.
"There are lots of ways to reduce the risk of a stress fracture, including vitamin D, calcium and weight-bearing exercise," Kim said in a statement. "But currently there is no protocol to predict in one's 20s or 30s who is likely to be at higher risk, and who should pursue these interventions before any sign of bone weakening."
With the growing number of people having their genomes assessed through direct-to-consumer genetic testing, he explained, such genetic risk scores may eventually offer "a relatively simple measure to identify those who should have their bone-mineral density tested and perhaps take steps at an early age to ensure their future bone health."
Starting with Affymetrix array-based genotyping data for 488,378 UK Biobank participants, Kim considered quantitative ultrasound-based BMD estimates, self-reported bone fracture histories, electronic health record data, and physical factors such as height, weight, and genotype-based sex. He ultimately focused on 394,929 individuals of European ancestry with available genotype and phenotype data for the BMD GWAS.
The analysis led to 142,417 SNPs with apparent ties to estimated BMD, including 1,362 variants with independent, genome-wide significant estimated BMD associations. The SNPs spanned 899 loci, with the strongest associations appearing at four sites on chromosomes 2, 6, and 7.
He compared the associations to those described previously for BMD or bone fracture, and brought in chromatin immunoprecipitation sequence data from ENCODE to narrow in on possible causal variants from the candidate SNP set. He also used several computational approaches to put together polygenic risk scores trained with data from the GWAS.
From there, Kim identified the algorithm with the most robust BMD prediction, which hinged on information from 22,886 SNPs. The predictive abilities of that algorithm got an additional boost when he incorporated height, weight, sex, and age. For example, he estimated that osteoporosis risk is more than 17 times higher in individuals with the lowest scores on this combined algorithm, dubbed the "BMD Osteoporosis Genetic" (BOG) risk score.
Kim cautioned that "the BOG algorithm identifies rare individuals at risk for low BMD, osteoporosis and fracture," though "its ability to discriminate cases from controls in the overall population is modest."