Kim-Anh Do

The paper identifies distinct microbial patterns and clusters associated with varying oral mucositis severity trajectories in head and neck cancer patients.

This study introduces AUC-PR as a better metric for evaluating in silico perturbation models' ability to predict differentially expressed genes, showing it outperforms R2 in capturing biologically relevant outcomes.

The paper introduces SMAGS, a regression framework that maximizes sensitivity at a given specificity for binary classification, showing 14% improvement in sensitivity on a colorectal cancer dataset.