Search

Medical research increasingly includes high-dimensional regression modeling with a need for error-in-variables methods. The Convex Conditioned Lasso (CoCoLasso) utilizes a reformulated Lasso objective function and an error-corrected cross-validation to enable error-in-variables regression, but requires heavy computations. Here, we develop a Block coordinate Descent Convex Conditioned Lasso (BDCoCoLasso) algorithm for modeling high-dimensional data that are only partially corrupted by measurement error. This algorithm separately optimizes the estimation of the uncorrupted and corrupted features in an iterative manner to reduce computational cost, with a specially calibrated formulation of cross-validation error. Through simulations, we show that the BDCoCoLasso algorithm successfully copes with much larger feature sets than CoCoLasso, and as expected, outperforms the naïve Lasso with enhanced estimation accuracy and consistency, as the intensity and complexity of measurement errors increase. Also, a new smoothly clipped absolute deviation penalization option is added that may be appropriate for some data sets. We apply the BDCoCoLasso algorithm to data selected from the UK Biobank. We develop and showcase the utility of covariate-adjusted genetic risk scores for body mass index, bone mineral density, and lifespan. We demonstrate that by leveraging more information than the naïve Lasso in partially corrupted data, the BDCoCoLasso may achieve higher prediction accuracy. These innovations, together with an R package, BDCoCoLasso, make error-in-variables adjustments more accessible for high-dimensional data sets. We posit the BDCoCoLasso algorithm has the potential to be widely applied in various fields, including genomics-facilitated personalized medicine research.

Some commonly prescribed drugs are associated with increased risk of osteoporotic fractures. However, fracture risk stratification using skeletal measures is not often performed to identify those at risk before these medications are prescribed. We tested whether a genomically predicted skeletal measure, speed of sound (gSOS) from heel ultrasound, which was developed in 341,449 individuals from UK Biobank and tested in a separate subset consisting of 80,027 individuals, is an independent risk factor for fracture in users of fracture-related drugs (FRDs). To do this, we first assessed 80,014 UK Biobank participants (including 12,678 FRD users) for incident major osteoporotic fracture (MOF, n = 1189) and incident hip fracture (n = 209). Effects of gSOS on incident fracture were adjusted for baseline clinical fracture risk factors. We found that each standard deviation decrease in gSOS increased the adjusted odds of MOF by 42% (95% confidence interval [CI] 1.34-1.51, p < 2 × 10-16 ) and of hip fracture by 31% (95% CI 1.15-1.50, p = 9 × 10-5 ). gSOS below versus above the mean increased the adjusted odds of MOF by 79% (95% CI 1.58-2.01, p < 2 × 10-16 ) and of hip fracture by 42% (95% CI 1.08-1.88, p = 1.3 × 10-2 ). Among FRD users, each standard deviation decrease in gSOS increased the adjusted odds of MOF by 29% (nMOF = 256, 95% CI 1.14-1.46, p = 7 × 10-5 ) and of hip fracture by 30% (nhip fracture = 68, 95% CI 1.02-1.65, p = 0.0335). FRD users with gSOS below versus above the mean had a 54% increased adjusted odds of MOF (95% 1.19-1.99, p = 8.95 × 10-4 ) and a twofold increased adjusted odds of hip fracture (95% 1.19-3.31, p = 8.5 × 10-3 ). We therefore showed that genomically predicted heel SOS is independently associated with incident fracture among FRD users. © 2020 American Society for Bone and Mineral Research.

Additional file 1: Fig. S1: Distribution of discrepancies in minor allele frequencies of SNPs between the China Kadoorie Biobank and the UK Biobank; Fig. S2: Distribution of standardized gSOS among 90,172 individuals of European ancestry; Fig. S3: Cumulative incidence of major osteoporotic fracture in (A) the MrOS US cohort and (B) the SOF cohort, and cumulative incidence of hip fracture in (C) the MrOS US cohort and (D) the SOF cohort; Table S1. Summary of model training and selection in the UK Biobank; Table S2. Summary of Pearson correlation between gSOS and clinical risk factors; Table S3. Summary of incidence of osteoporotic fracture; Table S4. Summary of predictive power of gSOS and clinical risk factors; Table S5. Details of net reclassification improvement of major osteoporotic fracture risk prediction using FRAX-gSOS; Table S6. Details of net reclassification improvement of hip fracture risk prediction using FRAX-gSOS.

Background Since screening programs identify only a small proportion of the population as eligible for an intervention, genomic prediction of heritable risk factors could decrease the number needing to be screened by removing individuals at low genetic risk. We therefore tested whether a polygenic risk score for heel quantitative ultrasound speed of sound (SOS)—a heritable risk factor for osteoporotic fracture—can identify low-risk individuals who can safely be excluded from a fracture risk screening program. Methods and findings A polygenic risk score for SOS was trained and selected in 2 separate subsets of UK Biobank (comprising 341,449 and 5,335 individuals). The top-performing prediction model was termed “gSOS”, and its utility in fracture risk screening was tested in 5 validation cohorts using the National Osteoporosis Guideline Group clinical guidelines (N = 10,522 eligible participants). All individuals were genome-wide genotyped and had measured fracture risk factors. Across the 5 cohorts, the average age ranged from 57 to 75 years, and 54% of studied individuals were women. The main outcomes were the sensitivity and specificity to correctly identify individuals requiring treatment with and without genetic prescreening. The reference standard was a bone mineral density (BMD)–based Fracture Risk Assessment Tool (FRAX) score. The secondary outcomes were the proportions of the screened population requiring clinical-risk-factor-based FRAX (CRF-FRAX) screening and BMD-based FRAX (BMD-FRAX) screening. gSOS was strongly correlated with measured SOS (r2 = 23.2%, 95% CI 22.7% to 23.7%). Without genetic prescreening, guideline recommendations achieved a sensitivity and specificity for correct treatment assignment of 99.6% and 97.1%, respectively, in the validation cohorts. However, 81% of the population required CRF-FRAX tests, and 37% required BMD-FRAX tests to achieve this accuracy. Using gSOS in prescreening and limiting further assessment to those with a low gSOS resulted in small changes to the sensitivity and specificity (93.4% and 98.5%, respectively), but the proportions of individuals requiring CRF-FRAX tests and BMD-FRAX tests were reduced by 37% and 41%, respectively. Study limitations include a reliance on cohorts of predominantly European ethnicity and use of a proxy of fracture risk. Conclusions Our results suggest that the use of a polygenic risk score in fracture risk screening could decrease the number of individuals requiring screening tests, including BMD measurement, while maintaining a high sensitivity and specificity to identify individuals who should be recommended an intervention., Vincenzo Forgetta and colleagues investigate possible utility of genetic information in assessing fracture risk., Author summary Why was this study done? Osteoporosis screening identifies only a small proportion of the screened population to be eligible for intervention. The prediction of heritable risk factors using polygenic risk scores could decrease the number of screened individuals by reassuring those with low genetic risk. We investigated whether the genetic prediction of heel quantitative ultrasound speed of sound (SOS)—a heritable risk factor for osteoporotic fracture—could be incorporated into an established screening guideline to identify individuals at low risk for osteoporosis. What did the researchers do and find? Using UK Biobank, we developed a polygenic risk score (gSOS) consisting of 21,717 genetic variants that was strongly correlated with SOS (r2 = 23.2%). Using the National Osteoporosis Guideline Group clinical assessment guidelines in 5 validation cohorts, we estimate that reassuring individuals with a high gSOS, rather than doing further assessments, could reduce the number of clinical-risk-factor-based Fracture Risk Assessment Tool (FRAX) tests and bone-density-measurement-based FRAX tests by 37% and 41%, respectively, while maintaining a high sensitivity and specificity to identify individuals who should be recommended an intervention. What do these findings mean? We show that genetic pre-screening could reduce the number of screening tests needed to identify individuals at risk of osteoporotic fractures. Therefore, the potential exists to improve the efficiency of osteoporosis screening programs without large losses in sensitivity or specificity to identify individuals who should receive an intervention. Further translational studies are needed to test the clinical applications of this polygenic risk score; however, our work shows how such scores could be tested in the clinic.

Vascular endothelial growth factor (VEGF) is important for bone formation and has been associated with osteoporosis in humans. Therefore, we conducted a two-sample Mendelian randomization study to test whether genetically decreased circulating VEGF was associated with decreased bone mineral density (BMD) and increased risk of fracture. Summary statistics from a genomewide association study (GWAS) meta-analysis of circulating VEGF level (n = 16,112) were used to identify 10 genetic variants explaining up to 52% of the variance in circulating VEGF levels. GWAS meta-analyses on dual-energy X-ray absorptiometry (DXA)-derived BMD of forearm, lumbar spine, and femoral neck (n = up to 32,735) and BMD estimated from heel calcaneus ultrasound (eBMD) (n = 426,824) were used to assess the effect of genetically lowered circulating VEGF levels on BMD. A GWAS meta-analysis including a total of 76,549 cases and 470,164 controls was used to assess the effect of genetically lowered circulating VEGF levels on risk of fracture. A natural log-transformed pg/mL decrease in circulating VEGF levels was not associated with a decrease in forearm BMD (0.02 standard deviation [SD], 95% confidence interval [CI] -0.024 to 0.064, p = 0.38), lumbar spine BMD (-0.005 SD, 95% CI -0.03 to 0.019, p = 0.67), femoral neck BMD (0.004 SD, 95% CI -0.017 to 0.026, p = 0.68), eBMD (-0.006 SD, 95% CI -0.012 to -0.001, p = 0.031) or risk of fracture (odds ratio = 0.99, 95% CI 0.98 to 1.0, p = 0.37) in inverse-variance-weighted Mendelian randomization analyses. Sensitivity analyses did not provide evidence that our results were influenced by pleiotropy. Genetically lowered circulating VEGF was not associated with a decrease in BMD or increased risk of fracture, suggesting that efforts to influence circulating VEGF level are unlikely to have beneficial effects on osteoporosis outcomes and that previous observational associations of circulating VEGF with BMD were influenced by confounding or reverse causation. © 2019 American Society for Bone and Mineral Research.

BackgroundGenomics-based prediction could be useful since genome-wide genotyping costs less than many clinical tests. We tested whether machine learning methods could provide a clinically-relevant genomic prediction of quantitative ultrasound speed of sound (SOS)—a risk factor for osteoporotic fracture.MethodsWe used 341,449 individuals from UK Biobank with SOS measures to develop genomically-predicted SOS (gSOS) using machine learning algorithms. We selected the optimal algorithm in 5,335 independent individuals and then validated it and its ability to predict incident fracture in an independent test dataset (N = 80,027). Finally, we explored whether genomic pre-screening could complement a UK-based osteoporosis screening strategy, based on the validated tool FRAX.ResultsgSOS explained 4.8-fold more variance in SOS than FRAX clinical risk factors (CRF) alone (r2 = 23% vs. 4.8%). A standard deviation decrease in gSOS, adjusting for the CRF-FRAX score was associated with a higher increased odds of incident major osteoporotic fracture (1,491 cases / 78,536 controls, OR = 1.91 [1.70-2.14], P = 10-28) than that for measured SOS (OR = 1.60 [1.50-1.69], P = 10-52) and femoral neck bone mineral density (147 cases / 4,594 controls, OR = 1.53 [1.27-1.83], P = 10-6). Individuals in the bottom decile of the gSOS distribution had a 3.25-fold increased risk of major osteoporotic fracture (P = 10-18) compared to the top decile. A gSOS-based FRAX score, identified individuals at high risk for incident major osteoporotic fractures better than the CRF-FRAX score (P = 10-14). Introducing a genomic pre-screening step into osteoporosis screening in 4,741 individuals reduced the number of required clinical visits from 2,455 to 1,273 and the number of BMD tests from 1,013 to 473, while only reducing the sensitivity to identify individuals eligible for therapy from 99% to 95%.InterpretationThe use of genotypes in a machine learning algorithm resulted in a clinically-relevant prediction of SOS and fracture, with potential to impact healthcare resource utilization.Research in ContextEvidence Before this StudyGenome-wide association studies have identified many loci associated with risk of clinically-relevant fracture risk factors, such as SOS. Yet, it is unclear if such information can be leveraged to identify those at risk for disease outcomes, such as osteoporotic fractures. Most previous attempts to predict disease risk from genotypes have used polygenic risk scores, which may not be optimal for genomic-prediction. Despite these obstacles, genomic-prediction could enable screening programs to be more efficient since most people screened in a population are not determined to have a level of risk that would prompt a change in clinical care. Genomic pre-screening could help identify individuals whose risk of disease is low enough that they are unlikely to benefit from screening.Added Value of this StudyUsing a large dataset of 426,811 individuals we trained and tested a machine learning algorithm to genomically-predict SOS. This metric, gSOS, had performance characteristics for predicting fracture risk that were similar to measured SOS and femoral neck BMD. Implementing a gSOS-based pre-screening step into the UK-based osteoporosis treatment guidelines reduced the number of individuals who would require screening clinical visits and skeletal testing by approximately 50%, while having little impact on the sensitivity to identify individuals at high risk for osteoporotic fracture.Implications of all of the Available EvidenceClinically-relevant genomic prediction of heritable traits is feasible using the machine learning algorithm presented here in large sample sizes. Genome-wide genotyping is now less expensive than many clinical tests, needs to be performed once over a lifetime and could risk stratify for multiple heritable traits and diseases years prior to disease onset, providing an opportunity for prevention. The implementation of such algorithms could improve screening efficiency, yet their cost-effectiveness will need to be ascertained in subsequent analyses.