| Application: | 33999, Identifying and quantifying risk factors for macular disorders using automated approaches to phenotyping |
| Title: | Chances and challenges of machine learning-based disease classification in genetic association studies illustrated on age-related macular degeneration |
| Size: | 428860 B |
| Archived: | 6 Apr 2021 |
| Stability: | Complete |
| Personal: | Contains individual-level data |
Return 3279
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
Notes
Imaging technology and machine learning algorithms for disease classification set the stage for high-throughput phenotyping and promising new avenues for genome-wide association studies (GWAS). To evaluate chances and challenges of utilizing a machine learning based disease classification in GWAS, we performed a study on fundus-image derived AMD in UK Biobank: we automatically classified fundus images based on a published neural network ensemble (images from 135,500 eyes; 68,400 persons) and performed a GWAS utilizing the derived any AMD phenotype. Predictions of machine learning algorithms can be erroneous and we quantified misclassification of automatically derived AMD in internal validation data with an additional manual AMD classification (gold standard; 4,001 eyes; 2,013 persons). We establish the utilization of a machine learning based phenotype in genetic association analyses as misclassification problem and developed a maximum likelihood approach (MLA) to account for misclassification when estimating genetic association. By combining a GWAS on automatically derived AMD and our MLA, we were able to dissect true association (ARMS2/HTRA1, CFH) from artefacts (near HERC2) and identified eye color as associated with the misclassification. On this example, we provide a proof-of-concept that a GWAS using machine learning derived disease classification yields relevant results and that misclassification needs to be considered in analysis. These findings generalize to other phenotypes and emphasize the utility of genetic data for understanding misclassification structure of machine learning algorithms.Application 33999
Identifying and quantifying risk factors for macular disorders using automated approaches to phenotyping
Therapeutic options for age-related macular degeneration (AMD) are limited and AMD is the leading cause of blindness in elderly. Disorders of the vitreoretinal interface require extensive surgical intervention. Both diseases are therefore linked to substantial individual and public health burden. We aim to contribute to improve the knowledge of the causes of these diseases. We approach this by using observational data to understand risk factors that increase the susceptibility for these diseases. These diseases can be diagnosed via fundus photography and optical coherence tomography (OCT). The UKBiobank data including data from these imaging techniques are ideal to investigate association between risk factors and diagnosis. One of the predominant challenges of these imaging data is the huge effort involved when manually analyzing these images. New automated approaches are available (deep learning). However, the performance of such automated approaches to yield high quality diagnosis compared to manual diagnosis is not known for these diseases. Specifically, we aim to apply deep learning algorithms for AMD and vitreoretinal boarder disorders that are already developed to UK Biobank data in order to evaluate the accuracy. These algorithms will be applied to the fundus photos and OCT images of UKBiobank. We will also manually grade images for a subgroup of persons comparing results with the automated diagnosis. We will analyze the association of genetic factors together with lifestyle and metabolic parameters with these retinal diseases. The expected Project Duration is one year. We expect our results to contribute to an improved understanding of the causes of retinal diseases, which will ultimately help to develop new and better therapeutic Options.
| Lead investigator: | Professor Iris Heid |
| Lead institution: | University of Regensburg |
1 Publication
| Pub ID | Title | Author(s) | Year | Journal |
|---|---|---|---|---|
| 3280 | Chances and challenges of machine learning-based disease classification in genetic association studies illustrated on age-related macular degeneration | Felix Guenther (+6) | 2020 | Genetic Epidemiology |
Enabling scientific discoveries that improve human health