Abstract
To characterize the genetic basis of chronic kidney disease, genome-wide association studies focused on chronic kidney disease-defining traits that were most readily measurable in large population samples. Today, convolutional neural network analysis of magnetic resonance imaging allows for accurate, large-scale kidney morphology assessment, enhancing research into chronic kidney disease determinants at population scale. Applying convolutional neural networks to UK Biobank data, Monteiro-Martins et al. generated genomic maps of kidney cortex, medulla, and sinus volumes, enabling more detailed research into kidney structure.