A Quantum Support Vector Machine based Sepsis Diagnosis Using Gene Expression Data

Analyzing differential gene expression in transcriptomic data provides crucial insights into molecular responses to specific biological conditions, potentially revealing valuable biomarkers derived from both established biological knowledge and data-driven approaches. In this study, we present a novel methodology for sepsis diagnosis utilizing immune-related gene expression data to identify optimal biomarker combinations. Our methodological framework incorporates multiple analytical steps: differential gene expression analysis, feature importance evaluation, and an integrated approach combining Recursive Feature Elimination (RFE) with Principal Component Analysis (PCA), and Quantum Support Vector Machine (QSVM)-based classification. The total of 41 immune-related genes is carefully selected to construct a comprehensive diagnostic panel. Implementation of QSVM classification yielded exceptional diagnostic performance, demonstrating superior results across all evaluation metrics. This approach represents a significant advancement in identifying reliable transcriptomic biomarkers for sepsis diagnosis and establishes a robust framework applicable to other complex diseases