PROPS

"PROPS" (PRObabilistic Pathway Score) is a groundbreaking pathway-based classification approach designed to address the limitations of traditional gene-based supervised machine learning models in distinguishing disease states and predicting disease progression. Recognizing the challenges posed by the sensitivity to noise and the lack of reproducibility in external validation sets of gene-based classifiers, especially in the context of complex, heterogeneous diseases, PROPS offers a robust solution by focusing on biological mechanisms rather than individual genes.

Unlike pathway-based classification methods relying solely on gene sets, PROPS innovatively incorporates gene interactions through probabilistic graphical models, allowing for a more accurate representation of the underlying biology and enhancing classification performance. The method calculates individualized pathway scores for classification, enabling the capture of diverse combinations of genes that contribute to the same phenotype.