Time Series Gene Expression Profiles Analysis Identified Several Potential Biomarkers for Sepsis

Sepsis is a life-threatening disorder and leads to organ dysfunction and death. Therefore, searching for more alternative biomarkers is of great significance for sepsis assessment and surveillance. In our study, the gene expression profiles of 163 samples from healthy controls and septic patients were analyzed and 8 gene co-expression modules were identified by constructing weighted gene co-expression network. The blue and yellow modules showed close correlations with the phenotypic trait "days postsepsis." Besides, differentially expressed genes (DEGs) over time in septic patients were screened using Short Time-series Expression Miner (STEM) program. The intersection of genes in the blue and yellow modules and DEGs, which were significantly enriched in "HTLV-1 infection" pathway, was analyzed with protein-protein interaction network. The logistic regression model based on these eight mRNAs was constructed to determine the type of the sample reliably. Eight vital genesCECR1,ANXA2,ELANE,CTSG,AZU1,PRTN3,LYZ, andDEFA4presented high scores and may be associated with sepsis, which provided candidate biomarkers for sepsis.