Dataset Information

An explainable prognostic prediction panel for sepsis based on serum amino acid profiles

ABSTRACT:

Sepsis is a life-threatening syndrome requiring aggressive management, and novel noninvasive biomarkers to enable risk stratification of sepsis with high confidence and to predict the sepsis-related outcomes are urgently needed.

A mass spectrometry–based quantitative method was used to analyze the abundance of serum amino acids between patients with sepsis and healthy controls (HC), in order to characterize alterations in amino acid profiles associated with sepsis. In addition, multiple machine learning methods were applied to construct a prognostic prediction model for patients with sepsis. The predictive performance was assessed and the feature contributions were screened, followed by the development of an explainable prognostic prediction panel for sepsis.

Sixty participants in the HC group and 172 patients in the sepsis group (82 patients with septic shock) were enrolled in this study. A discernible segregation trend in amino acid profiles between the HC and sepsis groups was disclosed, meanwhile the abundance of amino acids differed significantly among the HC, septic shock, and non-septic shock groups, indicating that amino acids could differentiate patients with sepsis from the HC group with good diagnostic performance. Then, 172 patients in the sepsis group were assigned to training and validation sets, and 130 patient patients with sepsis in external test set were collected. Five machine learning models (deephit, piecewise constant hazard, probability mass function, resource selection function, and extreme gradient boosting) were afterwards used and the deephit model was selected according to greater area under the curve and clinical benefits in the training, validation, and test sets. In the process of reducing features based on feature importance ranking, the Deephit model based on five features had the best ability to predict survival probability and an optimized Deephit model with screening five features including glutamine, glycine, lysine, pyroglutamic acid and proline was successfully developed to predict the prognostic risk probability for patients with sepsis.

INSTRUMENT(S): Liquid Chromatography MS - positive - hilic

PROVIDER: MTBLS14567 | MetaboLights | 2026-05-24

REPOSITORIES: MetaboLights

ACCESS DATA

Dataset's files

Source:

			Action	DRS
	a_MTBLS14567_LC-MS_positive_hilic.txt	Txt
	i_Investigation.txt	Txt
	m_MTBLS14567_LC-MS_positive_hilic_v2_maf.tsv	Tabular
	s_MTBLS14567.txt	Txt

Items per page:

1 - 4 of 4

Similar Datasets

Project description:Goal of the experiment: To identify correlated genes, pathways and groups of patients with systemic inflammatory response syndrome and septic shock that is indicative of biologically important processes active in these patients. Background: We measured gene expression levels and profiles of children with systemic inflammatory response syndrome (SIRS) and septic shock as a means for discovering patient sub-groups and gene signatures that are active in disease-affected individuals and potentially in patients with poor outcomes. Methods: Microarray and bioinformatics analyses of 123 microarray chips representing whole blood derived RNA from controls, children with SIRS, and children with septic shock. Results: A discovery-based filtering approach was undertaken to identify genes whose expression levels were altered in patients with SIRS or septic shock. Clustering of these genes identified 3 Major and several minor sub-groups of patients with SIRS or septic shock. The three groups differed with respect to incidence of septic shock and trended toward differences in mortality. Statistical analyses demonstrated that 6,435 gene probes were differentially regulated between the three patient sub-groups (false discovery rate < 0.001%). Of these gene probes, 623 gene probes within 7 major gene ontologies accounted for the majority of group differentiation. Network analyses of these 623 gene probes demonstrated 5 major gene networks that were differentially expressed between the 3 groups. Statistical comparison of septic shock survivors and non-survivors identified one major gene network that was under expressed in a high fraction of the non-survivors and identified potential biomarkers for poor outcome. Conclusions: This is the first genome-level demonstration of pediatric patient sub-groups with SIRS and septic shock. The sub-groups differ clinically and differentially express 5 major gene networks. We have identified gene signatures and potential biomarkers associated with poor outcome in children with septic shock. These data represent a major advancement in our genome-level understanding of pediatric SIRS and septic shock. Experiment Overall Design: Children < 10 years of age admitted to the pediatric intensive care unit and meeting the criteria for either SIRS or septic shock were eligible for the study. SIRS and septic shock were defined based on pediatric-specific criteria. We did not use separate categories of "sepsis" or "severe sepsis". Patients meeting criteria for "sepsis" or "severe sepsis" were placed in the categories of SIRS and septic shock, respectively, for study purposes. Control patients were recruited from the outpatient or inpatient departments of the participating institutions using the following exclusion criteria: a recent febrile illness (within 2 weeks), recent use of anti-inflammatory medications (within 2 weeks), or any history of chronic or acute disease associated with inflammation. Experiment Overall Design: After obtaining informed consent, blood samples were obtained on Day 1 of the study, and when possible on Day 3 of the study. Blood samples were divided for RNA extraction and isolation of serum. Severity of illness was calculated based on the PRISM III score. Organ failure was defined based on pediatric-specific criteria. Annotated clinical and laboratory data were collected daily while in the intensive care unit. Study patients were placed in the study categories of SIRS or Septic Shock on Day 1 of the study. On Day 3 of the study, patients were classified as SIRS, Septic Shock, or SIRS resolved (no longer meeting criteria for SIRS). All study patients were followed for 28 days to determine mortality or survival. Clinical, laboratory, and biological data were entered and stored using a web-based data base developed locally.

Dataset Information

An explainable prognostic prediction panel for sepsis based on serum amino acid profiles

Dataset's files

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets