Speaker
Description
High-dimensional molecular profiling has transformed cancer research by enabling detailed characterization of tumor biology and more personalized therapies. Yet patients with highly similar clinical and molecular profiles can still exhibit markedly different treatment responses, suggesting that important biological information remains hidden within these datasets.
This work is motivated by the hypothesis that biologically meaningful information may lie not only in individual molecular measurements, but also in the organization of patient populations. We investigate this hypothesis through complementary geometric and topological perspectives,
asking whether patient organization provides biological insight beyond individual molecular features.
To investigate this question, we employ a complementary geometric and topological framework to characterize relationships and connectivity among patients, rather than relying solely on low-dimensional representations or discrete patient groups. Using breast cancer proteomic data from a clinical trial, we show how this perspective provides a richer characterization of molecular heterogeneity by revealing organizational features not readily captured through conventional analyses.
More broadly, this work highlights the value of studying patient organization as a complementary perspective for understanding high-dimensional molecular data and demonstrates how geometric and topological analyses jointly reveal meaningful biological structure.
