Related Experiment Video
Updated: Sep 16, 2026

A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
Identification of potential key genes involved in iron deficiency for sepsis: A retrospective cohort and
Zehong Wu1, Zhangqing Yi2,3, Hanyi Yao1
1Department of Cardiovascular Surgery, The Third Xiangya Hospital of Central South University, Changsha, Hunan, China.
Abstract:
Iron overload has been associated with sepsis, but the role of iron deficiency and its molecular links remain unclear. We investigated the association between iron deficiency and sepsis and identified candidate genes potentially linking these conditions. MIMIC-IV data were used to assess the association between serum iron and sepsis status. Transcriptomic datasets from dietary iron-deficient mice (GSE10421), LPS-induced septic mice (GSE267388), and a human blood sepsis cohort (GSE137340) were sequentially analyzed to identify and externally evaluate candidate genes. IEU Open GWAS summary statistics were used for exploratory Mendelian randomization (MR). Exploratory drug prediction was performed using L1000FWD, followed by molecular docking analysis. Patients with sepsis had significantly lower serum iron levels, and restricted cubic spline analysis showed a nonlinear association between serum iron and the odds of sepsis. Cross-tissue transcriptomic analysis identified Sqle, Lss, and Rdh11 as candidate genes. In the human blood cohort, SQLE and RDH11 were significantly increased, whereas LSS was not significantly altered. Exploratory MR showed that genetically proxied SQLE expression was associated with higher odds of sepsis (odds ratio [OR] = 1.23, P = 1.67 × 10-3), whereas LSS expression was associated with lower odds (OR = 0.97, P = 8.90 × 10-4); RDH11 showed no significant association (OR = 1.01, P = .90). Drug prediction identified ML106 as the top-ranked candidate drug, and molecular docking predicted potential binding poses with SQLE and LSS. Serum iron showed a nonlinear association with sepsis status. SQLE, LSS, and RDH11 emerged as candidate genes, with concordant expression changes of SQLE and RDH11 observed in human blood. MR findings for SQLE and LSS were exploratory and require further validation. ML106 was identified through exploratory drug prediction and requires experimental validation before its therapeutic relevance can be established.