Background: The microvasculature, the smallest blood vessels in the body, has key roles in maintenance of organ health and tumorigenesis. The retinal fundus is a window for human in vivo noninvasive assessment of the microvasculature. Large-scale complementary machine learning-based assessment of the retinal vasculature with phenome-wide and genome-wide analyses may yield new insights into human health and disease. Methods: We used 97 895 retinal fundus images from 54 813 UK Biobank participants. Using convolutional neural networks to segment the retinal microvasculature, we calculated vascular density and fractal dimension as a measure of vascular branching complexity. We associated these indices with 1866 incident International Classification of Diseases –based conditions (median 10-year follow-up) and 88 quantitative traits, adjusting for age, sex, smoking status, and ethnicity. Results: Low retinal vascular fractal dimension and density were significantly associated with higher risks for incident mortality, hypertension, congestive heart failure, renal failure, type 2 diabetes, sleep apnea, anemia, and multiple ocular conditions, as well as corresponding quantitative traits. Genome-wide association of vascular fractal dimension and density identified 7 and 13 novel loci, respectively, that were enriched for pathways linked to angiogenesis (eg, vascular endothelial growth factor, platelet-derived growth factor receptor, angiopoietin, and WNT signaling pathways) and inflammation (eg, interleukin, cytokine signaling). Conclusions: Our results indicate that the retinal vasculature may serve as a biomarker for future cardiometabolic and ocular disease and provide insights into genes and biological pathways influencing microvascular indices. Moreover, such a framework highlights how deep learning of images can quantify an interpretable phenotype for integration with electronic health record, biomarker, and genetic data to inform risk prediction and risk modification.

Background - Both lifestyle and genetic factors confer risk for cardiovascular diseases, type 2 diabetes (T2D), and dyslipidemia. However, the interactions between these two groups of risk factors were not comprehensively understood due to previous poor estimation of genetic risk. Here we set out to develop enhanced polygenic risk scores (PRS), and systematically investigate multiplicative and additive interactions between PRS and lifestyle for coronary artery disease, atrial fibrillation, T2D, total cholesterol, triglyceride, and LDL-cholesterol. Methods - Our study included 276,096 unrelated white British participants from the UK Biobank. We investigated several PRS methods (P+T, LDpred, PRS-CS, and AnnoPred), and showed that AnnoPred achieved consistently improved prediction accuracy for all six diseases/traits. With enhanced PRS and combined lifestyle status categorized by smoking, body mass index, physical activity, and diet, we investigated both multiplicative and additive interactions between PRS and lifestyle using regression models. Results - We observed that healthy lifestyle reduced disease incidence by similar multiplicative magnitude across different PRS groups. The absolute risk reduction (ARR) from lifestyle adherence was however significantly greater in individuals with higher PRS. Specifically, for T2D, the ARR from lifestyle adherence was 12.4% (95% CI, 10.0%-14.9%) in the top 1% PRS versus 2.8% (95% CI, 2.3%-3.3%) in the bottom PRS decile, leading to a ratio of more than 4.4. We also observed a significant interaction effect between PRS and lifestyle on triglyceride level. Conclusions - By leveraging functional annotations, AnnoPred outperforms state-of-the-art methods on quantifying genetic risk through PRS. Our analyses based on enhanced PRS suggest that individuals with high genetic risk may derive similar relative but greater absolute benefit from lifestyle adherence.

Cancers harbor many somatic mutations and germline variants, we hypothesized that the combined effect of germline variants that alter the structure, expression, or function of protein-coding regions of cancer-biology related genes (gHFI) determines which and how many somatic mutations (sM) must occur for malignant transformation. We show that gHFI and sM affect overlapping genes and the average number of gHFI in cancer hallmark genes is higher in patients who develop cancer at a younger age (r = −0.77, P = 0.0051), while the average number of sM increases in increasing age groups (r = 0.92, P = 0.000073). A strong negative correlation exists between average gHFI and average sM burden in increasing age groups (r = −0.70, P = 0.017). In early-onset cancers, the larger gHFI burden in cancer genes suggests a greater contribution of germline alterations to the transformation process while late-onset cancers are more driven by somatic mutations.

Genetic correlation is the correlation of phenotypic effects by genetic variants across the genome on two phenotypes. It is an informative metric to quantify the overall genetic similarity between complex traits, which provides insights into their polygenic genetic architecture. Several methods have been proposed to estimate genetic correlation based on data collected from genome-wide association studies (GWAS). Due to the easy access of GWAS summary statistics and computational efficiency, methods only requiring GWAS summary statistics as input have become more popular than methods utilizing individual-level genotype data. Here, we present a benchmark study for different summary-statistics-based genetic correlation estimation methods through simulation and real data applications. We focus on two major technical challenges in estimating genetic correlation: marker dependency caused by linkage disequilibrium (LD) and sample overlap between different studies. To assess the performance of different methods in the presence of these two challenges, we first conducted comprehensive simulations with diverse LD patterns and sample overlaps. Then we applied these methods to real GWAS summary statistics for a wide spectrum of complex traits. Based on these experiments, we conclude that methods relying on accurate LD estimation are less robust in real data applications due to the imprecision of LD obtained from reference panels. Our findings offer guidance on how to choose appropriate methods for genetic correlation estimation in post-GWAS analysis.