A robust and generalizable transcriptomic signature for sepsis diagnostics

Motivation: High-throughput sequencing can detect tens of thousands of genes in parallel, providing opportunities for improving the diagnostic accuracy of multiple diseases including sepsis, which is an aggressive inflammatory response to infection that can cause organ failure and death. Early screening of sepsis is essential in clinic, but no effective diagnostic biomarkers are available yet. Results: We present a novel method, Recurrent Logistic Regression (RLR), to identify diagnostic biomarkers for sepsis from the blood transcriptome data. Immune-related gene expression profiles are studied among 1,144 sepsis samples and 240 normal samples across three detection platforms, including Affymetrix Human Genome U133 Plus 2.0, Affymetrix Human Genome U219 Array, and Agilent Human Gene Expression 4x44K v2 Microarray. A panel including five genes, LRRN3, IL2RB, FCER1A, TLR5, and S100A12, are determined as diagnostic biomarkers (LIFTS) for sepsis on two discovery cohorts. LIFTS discriminates patients with sepsis from normal controls in high accuracy (AUROC = 0.9959 on average; IC = [0.9722-1.0]) on nine validation cohorts across three independent platforms, which outperforms existing markers. Each individual gene has certain distinguishability for the detection of sepsis but cannot achieve as high performance as the entire gene panel. Network and functional analysis illustrated that the interactors of the five genes are closely interacted and are commonly involved in several biological functions, including growth hormone synthesis, chemokine signaling pathway, B cell receptor signaling pathway, etc.