AI可识别心脏病相关基因并预测潜在药物

Group 1 - The core achievement is the development of an AI tool named CardioKG, which generates detailed views of heart structure and function, identifies previously unknown genes related to heart diseases, and predicts two potential therapeutic drugs [1] - The research utilized cardiac imaging data from 4,280 patients with atrial fibrillation, heart failure, or heart attacks, and 5,304 healthy participants, extracting over 200,000 image-based features to train the model [1] - CardioKG successfully predicted gene-disease associations and explored drug repurposing possibilities, identifying multiple previously unknown heart disease-related genes and two potential heart disease treatments: methotrexate for heart failure and GLP-1 agonists for atrial fibrillation [1] Group 2 - The application prospects of CardioKG extend beyond cardiology, with potential adaptations for research on other organs and tissues, such as the brain and body fat, supporting the exploration of new therapies for diseases like dementia and obesity [2]