Ranking Biomarkers Through Mutual Information
This work addresses a key problem in biomarker discovery for clinical researchers, offering incremental improvements through efficient estimators and visualization.
The paper tackles the challenge of disentangling predictive and prognostic biomarkers in clinical trials by formalizing biomarker ranking as an optimization of information theoretic quantities, and introduces a new visualization tool to capture their strengths.
We study information theoretic methods for ranking biomarkers. In clinical trials there are two, closely related, types of biomarkers: predictive and prognostic, and disentangling them is a key challenge. Our first step is to phrase biomarker ranking in terms of optimizing an information theoretic quantity. This formalization of the problem will enable us to derive rankings of predictive/prognostic biomarkers, by estimating different, high dimensional, conditional mutual information terms. To estimate these terms, we suggest efficient low dimensional approximations, and we derive an empirical Bayes estimator, which is suitable for small or sparse datasets. Finally, we introduce a new visualisation tool that captures the prognostic and the predictive strength of a set of biomarkers. We believe this representation will prove to be a powerful tool in biomarker discovery.