Non-linear machine learning models to improve variant prediction among clinical breast cancer patients

Presenting Author: Linda M. Polfus, PhD

Take home points: 

1. Machine Learning approaches capturing common variants interacting with rare Pathogenic Variants in high-risk genes create a ‘Polygenic Risk Score’
2. Beyond population baseline risk, genetics and age contribute to breast cancer risk, useful to predict which patients are at highest risk and benefit most from screening or certain drugs (i.e., platinum therapy or PARP inhibitors)