Skip to content

Predicting kidney stone risk using CNNs on genetic data, analyzing 400+ SNPs for precise risk stratification with Polygenic Risk Scores (PRS)

Notifications You must be signed in to change notification settings

amrsalm/deep-learning-prs

Repository files navigation

Kidney Stone Risk Prediction Using CNNs

This project leverages deep learning, specifically Convolutional Neural Networks (CNNs), to predict susceptibility to kidney stones based on genotypic data. We process genetic variants from over 400 single nucleotide polymorphisms (SNPs) and use these to generate Polygenic Risk Scores (PRS) for a cohort of 60 individuals. The CNN model is designed to capture complex patterns in the genetic data to distinguish between high-risk and low-risk individuals effectively. Evaluation metrics include accuracy, ROC-AUC, precision, and recall to assess model performance.

Note: This project is in the very early stages and is part of a senior thesis at Case Western Reserve University (CWRU).

About

Predicting kidney stone risk using CNNs on genetic data, analyzing 400+ SNPs for precise risk stratification with Polygenic Risk Scores (PRS)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages