Age-related macular degeneration (AMD) is a common threat to vision. While classification of disease stages is critical to understanding disease risk and progression, several systems based on color ...fundus photographs are known. Most of these require in-depth and time-consuming analysis of fundus images. Herein, we present an automated computer-based classification algorithm.
Algorithm development for AMD classification based on a large collection of color fundus images. Validation is performed on a cross-sectional, population-based study.
We included 120 656 manually graded color fundus images from 3654 Age-Related Eye Disease Study (AREDS) participants. AREDS participants were >55 years of age, and non-AMD sight-threatening diseases were excluded at recruitment. In addition, performance of our algorithm was evaluated in 5555 fundus images from the population-based Kooperative Gesundheitsforschung in der Region Augsburg (KORA; Cooperative Health Research in the Region of Augsburg) study.
We defined 13 classes (9 AREDS steps, 3 late AMD stages, and 1 for ungradable images) and trained several convolution deep learning architectures. An ensemble of network architectures improved prediction accuracy. An independent dataset was used to evaluate the performance of our algorithm in a population-based study.
κ Statistics and accuracy to evaluate the concordance between predicted and expert human grader classification.
A network ensemble of 6 different neural net architectures predicted the 13 classes in the AREDS test set with a quadratic weighted κ of 92% (95% confidence interval, 89%–92%) and an overall accuracy of 63.3%. In the independent KORA dataset, images wrongly classified as AMD were mainly the result of a macular reflex observed in young individuals. By restricting the KORA analysis to individuals >55 years of age and prior exclusion of other retinopathies, the weighted and unweighted κ increased to 50% and 63%, respectively. Importantly, the algorithm detected 84.2% of all fundus images with definite signs of early or late AMD. Overall, 94.3% of healthy fundus images were classified correctly.
Our deep learning algoritm revealed a weighted κ outperforming human graders in the AREDS study and is suitable to classify AMD fundus images in other datasets using individuals >55 years of age.
The growing sample size of genome-wide association studies has facilitated the discovery of gene-environment interactions (GxE). Here we propose a maximum likelihood method to estimate the ...contribution of GxE to continuous traits taking into account all interacting environmental variables, without the need to measure any. Extensive simulations demonstrate that our method provides unbiased interaction estimates and excellent coverage. We also offer strategies to distinguish specific GxE from general scale effects. Applying our method to 32 traits in the UK Biobank reveals that while the genetic risk score (GRS) of 376 variants explains 5.2% of body mass index (BMI) variance, GRSxE explains an additional 1.9%. Nevertheless, this interaction holds for any variable with identical correlation to BMI as the GRS, hence may not be GRS-specific. Still, we observe that the global contribution of specific GRSxE to complex traits is substantial for nine obesity-related measures (including leg impedance and trunk fat-free mass).
The R package EasyStrata facilitates the evaluation and visualization of stratified genome-wide association meta-analyses (GWAMAs) results. It provides (i) statistical methods to test and account for ...between-strata difference as a means to tackle gene-strata interaction effects and (ii) extended graphical features tailored for stratified GWAMA results. The software provides further features also suitable for general GWAMAs including functions to annotate, exclude or highlight specific loci in plots or to extract independent subsets of loci from genome-wide datasets. It is freely available and includes a user-friendly scripting interface that simplifies data handling and allows for combining statistical and graphical functions in a flexible fashion.
EasyStrata is available for free (under the GNU General Public License v3) from our Web site www.genepi-regensburg.de/easystrata and from the CRAN R package repository cran.r-project.org/web/packages/EasyStrata/.
Supplementary data are available at Bioinformatics online.
Genome-wide association meta-analyses (GWAMAs) conducted separately by two strata have identified differences in genetic effects between strata, such as sex-differences for body fat distribution. ...However, there are several approaches to identify such differences and an uncertainty which approach to use. Assuming the availability of stratified GWAMA results, we compare various approaches to identify between-strata differences in genetic effects. We evaluate type I error and power via simulations and analytical comparisons for different scenarios of strata designs and for different types of between-strata differences. For strata of equal size, we find that the genome-wide test for difference without any filtering is the best approach to detect stratum-specific genetic effects with opposite directions, while filtering for overall association followed by the difference test is best to identify effects that are predominant in one stratum. When there is no a priori hypothesis on the type of difference, a combination of both approaches can be recommended. Some approaches violate type I error control when conducted in the same data set. For strata of unequal size, the best approach depends on whether the genetic effect is predominant in the larger or in the smaller stratum. Based on real data from GIANT (>175 000 individuals), we exemplify the impact of the approaches on the detection of sex-differences for body fat distribution (identifying up to 10 loci). Our recommendations provide tangible guidelines for future GWAMAs that aim at identifying between-strata differences. A better understanding of such effects will help pinpoint the underlying mechanisms.
The problem of the genetics of related phenotypes is often addressed by analyzing adjusted-model traits, but such traits warrant cautious interpretation. Here, we adopt a joint view of adiposity ...traits in ~322,154 subjects (GIANT consortium). We classify 159 signals associated with body mass index (BMI), waist-to-hip ratio (WHR), or WHR adjusted for BMI (WHRadjBMI) at P < 5 × 10
, into four classes based on the direction of their effects on BMI and WHR. Our classes help differentiate adiposity genetics with respect to anthropometry, fat depots, and metabolic health. Class-specific Mendelian randomization reveals that variants associated with both WHR-decrease and BMI increase are linked to metabolically rather favorable adiposity through beneficial hip fat. Class-specific enrichment analyses implicate digestive systems as a pathway in adiposity genetics. Our results demonstrate that WHRadjBMI variants capture relevant effects of "unexpected fat distribution given the BMI" and that a joint view of the genetics underlying related phenotypes can inform on important biology.
To systematically analyze thicknesses of retinal layers in an older population and their link to early age-related macular degeneration (AMD).
In the AugUR baseline survey from a population aged ≥70 ...years, we conducted multimodal retinal imaging, including spectral-domain optical coherence tomography. Autosegmentation of eight distinct retinal layers was followed by manual correction of segmentation errors. AMD status was graded on color fundus images according to the Three Continent AMD Consortium Severity Scale. We tested the association of early AMD on retinal layer thicknesses by using linear mixed models and replicated significant results in independent data also from the AugUR platform.
When comparing layer thicknesses between early AMD and no AMD (822 eyes, 449 participants), the retinal pigment epithelium/Bruch's membrane complex demonstrated a statistically significant thickening (e.g., P = 6.41 × 10-92 for severe early versus no AMD) and photoreceptor layers showed a significant thinning. Autosegmented retinal layer thicknesses revealed similar associations as manually corrected values but underestimated some effects. Independent replication analysis in 1026 eyes (546 participants) confirmed associations (e.g., P = 9.38 × 10-36 for retinal pigment epithelium/Bruch's membrane complex, severe early versus no AMD).
This first population-based study on spectral-domain optical coherence tomography-derived retinal layer thicknesses in a total of ∼1000 individuals provides insights into the reliability of autosegmentation and layer-specific reference values for an older population. Our findings show a difference in thicknesses between early AMD and no AMD for some retinal layers, suggesting these as potential imaging biomarkers. The thinning of photoreceptor layers substantiates a photoreceptor cell loss/damage already occurring in early AMD.
Genome-wide association studies (GWAS) have identified hundreds of genetic loci associated with kidney function. By combining these findings with post-GWAS information (e.g., statistical fine-mapping ...to identify independent association signals and to narrow down signals to causal variants; or different sources of annotation data), new hypotheses regarding physiology and disease aetiology can be obtained. These hypotheses need to be tested in laboratory experiments, for example, to identify new therapeutic targets. For this purpose, the evidence obtained from GWAS and post-GWAS analyses must be processed and presented in a way that they are easily accessible to kidney researchers without specific GWAS expertise. With the focus on kidney function related evidence, KidneyGPS fills a gap between large general platforms for accessing GWAS and post-GWAS results and the specific needs of the kidney research community. This makes KidneyGPS an important platform for kidney researchers to help translate in silico research results into in vitro or in vivo research.
Dilated cardiomyopathy (DCM) is a structural heart disease with strong genetic background. Monogenic forms of DCM are observed in families with mutations located mostly in genes encoding structural ...and sarcomeric proteins. However, strong evidence suggests that genetic factors also affect the susceptibility to idiopathic DCM. To identify risk alleles for non-familial forms of DCM, we carried out a case-control association study, genotyping 664 DCM cases and 1,874 population-based healthy controls from Germany using a 50K human cardiovascular disease bead chip covering more than 2,000 genes pre-selected for cardiovascular relevance. After quality control, 30,920 single nucleotide polymorphisms (SNP) were tested for association with the disease by logistic regression adjusted for gender, and results were genomic-control corrected. The analysis revealed a significant association between a SNP in HSPB7 gene (rs1739843, minor allele frequency 39%) and idiopathic DCM (p = 1.06 × 10⁻⁶, OR = 0.67 95% CI 0.57-0.79 for the minor allele T). Three more SNPs showed p < 2.21 × 10⁻⁵. De novo genotyping of these four SNPs was done in three independent case-control studies of idiopathic DCM. Association between SNP rs1739843 and DCM was significant in all replication samples: Germany (n =564, n = 981 controls, p = 2.07 × 10⁻³, OR = 0.79 95% CI 0.67-0.92), France 1 (n = 433 cases, n = 395 controls, p =3.73 × 10⁻³, OR = 0.74 95% CI 0.60-0.91), and France 2 (n = 249 cases, n = 380 controls, p = 2.26 × 10⁻⁴, OR = 0.63 95% CI 0.50-0.81). The combined analysis of all four studies including a total of n = 1,910 cases and n = 3,630 controls showed highly significant evidence for association between rs1739843 and idiopathic DCM (p = 5.28 × 10⁻¹³, OR= 0.72 95% CI 0.65-0.78). None of the other three SNPs showed significant results in the replication stage.This finding of the HSPB7 gene from a genetic search for idiopathic DCM using a large SNP panel underscores the influence of common polymorphisms on DCM susceptibility.
Abstract Background Micro- and macrovascular diseases are common in patients with type 2 diabetes mellitus (T2D) and may be partly caused by nocturnal hypoxemia. The study aimed to characterize the ...composition of nocturnal hypoxemic burden and to assess its association with micro- and macrovascular disease in patients with T2D. Methods This cross-sectional analysis includes overnight oximetry from 1247 patients with T2D enrolled in the DIACORE (DIAbetes COhoRtE) study. Night-time spent below a peripheral oxygen saturation of 90% (T90) as well as T90 associated with non-specific drifts in oxygen saturation (T90 non − specific ), T90 associated with acute oxygen desaturation (T90 desaturation ) and desaturation depths were assessed. Binary logistic regression analyses adjusted for known risk factors (age, sex, smoking status, waist-hip ratio, duration of T2D, HbA1c, pulse pressure, low-density lipoprotein, use of statins, and use of renin-angiotensin-aldosterone system inhibitors) were used to assess the associations of such parameters of hypoxemic burden with chronic kidney disease (CKD) as a manifestation of microvascular disease and a composite of cardiovascular diseases (CVD) reflecting macrovascular disease. Results Patients with long T90 were significantly more often affected by CKD and CVD than patients with a lower hypoxemic burden (CKD 38% vs. 28%, p < 0.001; CVD 30% vs. 21%, p < 0.001). Continuous T90 desaturation and desaturation depth were associated with CKD (adjusted OR 1.01 per unit, 95% CI 1.00; 1.01, p = 0.008 and OR 1.30, 95% CI 1.06; 1.61, p = 0.013, respectively) independently of other known risk factors for CKD. For CVD there was a thresholdeffect, and only severly and very severly increased T90 non−specific was associated with CVD (Q3;Q4 versus Q1;Q2, adjusted OR 1.51, 95% CI 1.12; 2.05, p = 0.008) independently of other known risk factors for CVD. Conclusion While hypoxemic burden due to oxygen desaturations and the magnitude of desaturation depth were significantly associated with CKD, only severe hypoxemic burden due to non-specific drifts was associated with CVD. Specific types of hypoxemic burden may be related to micro- and macrovascular disease.
Genes underneath signals from genome-wide association studies (GWAS) for kidney function are promising targets for functional studies, but prioritizing variants and genes is challenging. By GWAS ...meta-analysis for creatinine-based estimated glomerular filtration rate (eGFR) from the Chronic Kidney Disease Genetics Consortium and UK Biobank (n = 1,201,909), we expand the number of eGFRcrea loci (424 loci, 201 novel; 9.8% eGFRcrea variance explained by 634 independent signal variants). Our increased sample size in fine-mapping (n = 1,004,040, European) more than doubles the number of signals with resolved fine-mapping (99% credible sets down to 1 variant for 44 signals, ≤5 variants for 138 signals). Cystatin-based eGFR and/or blood urea nitrogen association support 348 loci (n = 460,826 and 852,678, respectively). Our customizable tool for Gene PrioritiSation reveals 23 compelling genes including mechanistic insights and enables navigation through genes and variants likely relevant for kidney function in human to help select targets for experimental follow-up.
PDF
