AI-Powered Model Aims to Decode Genetic Variants in Stargardt Disease
核心洞察
Researchers are developing an AI model to classify genetic variants of the ABCA4 (搜索) gene, crucial for understanding Stargardt disease (搜索).
The model integrates genetic, clinical, and structural data to predict the pathogenicity of previously unclassified variants.
This research aims to personalize medicine by identifying appropriate clinical trials and therapies for individuals with Stargardt disease (搜索).
An interdisciplinary research team at the University of Delaware is developing a machine-learning model to better understand the genetic drivers of inherited retinal diseases, particularly Stargardt disease (搜索). The research, led by Esther Biswas-Fiss, aims to classify the pathogenicity of variants in the ABCA4 (搜索) gene, which is responsible for Stargardt disease. This initiative is supported by a $2.3 million grant from the National Institutes of Health's National Eye Institute (NIH/NEI).
Stargardt disease (搜索), a rare genetic eye disorder, affects approximately 30,000 individuals in the United States. The disease is caused by mutations in the ABCA4 (搜索) gene, leading to progressive vision loss. However, the clinical presentation and severity of Stargardt disease can vary significantly among patients, making diagnosis and treatment challenging. According to Biswas-Fiss, the variable expressivity of the disease means that "every patient is different and doesn’t present the same clinically," underscoring the need to understand the genetic drivers of inherited blindness.
Addressing Unclassified Genetic Variants
Nearly 40% of the ABCA4 (搜索)'s genetic variants remain unclassified in pathogenicity, posing significant challenges for accurate diagnosis and treatment. The research team, including Sam Biswas, Shawn Polson, and Barry Bodt, is employing artificial intelligence to address this issue. They are developing a machine-learning model that analyzes genetic, clinical, and structural data to predict the pathogenicity of these variants.
"Through this NIH grant, we will develop a model that looks at all variants that have been unequivocally classified as pathogenic and benign to create a data set that can use algorithms to make predictions," Biswas-Fiss explained. The model will leverage 3D computational modeling to analyze the protein structures of these variants and integrate clinical data, such as the frequency of these variants in the population.
Personalized Medicine and Clinical Trials
The ultimate goal of this research is to enable more personalized medicine for individuals with Stargardt disease (搜索). By understanding the nature and pathogenicity of specific ABCA4 (搜索) variants, clinicians can better determine appropriate clinical trials and precision therapies for patients. Biswas-Fiss emphasized that this research will "help determine what clinical trials or precision therapies might be appropriate" and "assist them with life-planning decisions."
This project builds upon previous funding received by Biswas-Fiss, including a $300,000 grant from the Foundation Fighting Blindness. The ongoing research also provides opportunities for underrepresented students in science, such as medical sciences doctoral candidate Senem Cevik, who has worked alongside Biswas-Fiss on this research for several years.