Dataset Information

Machine learning identifies clinical sepsis phenotypes that translate to the plasma proteome

ABSTRACT: Background: Sepsis therapy is still limited to the treatment of the underlying infection and supportive measures. Therapeutic options that address the molecular changes of sepsis have not yet been identified. With the aim of a future individualized therapy, we used unsupervised machine learning (ML) to identify clinical phenotypes in a prospective multicenter cohort of patients with sepsis and characterized them using plasma proteomics. Methods: Routine clinical data and blood samples were collected from 384 sepsis patients. Clinical phenotypes were identified based on clinical routine measurements using the k-means algorithm. In addition, plasma samples from 276 patients were analyzed using liquid chromatography-coupled tandem mass spectrometry (LC-MS/MS). The obtained data were analyzed and interpreted in the context of the phenotypes and supervised ML classifiers were developed to prospectively allocate patients to the clusters and to identify the most important features for discrimination of the phenotypes. Results: Three clinical phenotypes were identified. Cluster C was characterized by the highest disease severity and multi-organ failure with leading liver failure and a mortality rate of 92%. Cluster B, with a mortality rate of 45 %, also showed relevant organ failure, with renal damage being particularly prominent in comparison to cluster A. The plasma proteome reflected the clinical features of the phenotypes and revealed the excessive consumption of complement and coagulation factors in severe sepsis. Supervised ML and feature importance analysis underlined these findings and highlighted specific clinical measures and proteins. Conclusions: ML identified clinical phenotypes showed different degrees of sepsis severity and could be translated to the plasma proteome. Plasma proteomics provided novel insights into the molecular processes of sepsis and allowed the characterization of the phenotypes. This approach represents a blueprint to identify molecular features of sepsis subgroups and may pave the way for future targeted sepsis therapy.

INSTRUMENT(S):

ORGANISM(S): Homo Sapiens (human)

TISSUE(S): Blood Plasma

DISEASE(S): Bacterial Sepsis,Severe Covid-19

SUBMITTER: Thilo Bracht

LAB HEAD: Thilo Bracht

PROVIDER: PXD058562 | Pride | 2025-12-08

REPOSITORIES: Pride

ACCESS DATA

Dataset's files

Source:

Items per page:

1 - 5 of 286

Publications

Machine learning identifies clinical sepsis phenotypes that translate to the plasma proteome.

Bracht Thilo T Weber Maike M Kappler Kerstin K Palmowski Lars L Bayer Malte M Schork Karin K Rahmel Tim T Unterberg Matthias M Haberl Helge H Wolf Alexander A Koos Björn B Rump Katharina K Ziehe Dominik D Limper Ulrich U Henzler Dietrich D Ehrentraut Stefan Felix SF von Groote Thilo T Zarbock Alexander A Marcus-Alic Katrin K Eisenacher Martin M Adamzik Michael M Sitek Barbara B Nowak Hartmuth H

Infection 20250828 6

<h4>Background</h4>Sepsis therapy is still limited to treatment of the underlying infection and supportive measures. To date, various sepsis subtypes were proposed, but therapeutic options addressing the molecular changes of sepsis were not identified. With the aim of a future individualized therapy, we used machine learning (ML) to identify clinical phenotypes and their temporal development in a prospective, multicenter sepsis cohort and characterized them using plasma proteomics.<h4>Methods</h ...[more]

PMID: 40875167

Similar Datasets

Project description:Background: Sepsis therapy is still limited to the treatment of the underlying infection and supportive measures. Therapeutic options that address the molecular changes of sepsis have not yet been identified. With the aim of a future individualized therapy, we used unsupervised machine learning (ML) to identify clinical phenotypes in a prospective multicenter cohort of patients with sepsis and characterized them using plasma proteomics. Methods: Routine clinical data and blood samples were collected from 384 sepsis patients. Clinical phenotypes were identified based on clinical routine measurements using the k-means algorithm. In addition, plasma samples from 276 patients were analyzed using liquid chromatography-coupled tandem mass spectrometry (LC-MS/MS). The obtained data were analyzed and interpreted in the context of the phenotypes and supervised ML classifiers were developed to prospectively allocate patients to the clusters and to identify the most important features for discrimination of the phenotypes. Results: Three clinical phenotypes were identified. Cluster C was characterized by the highest disease severity and multi-organ failure with leading liver failure and a mortality rate of 92%. Cluster B, with a mortality rate of 45 %, also showed relevant organ failure, with renal damage being particularly prominent in comparison to cluster A. The plasma proteome reflected the clinical features of the phenotypes and revealed the excessive consumption of complement and coagulation factors in severe sepsis. Supervised ML and feature importance analysis underlined these findings and highlighted specific clinical measures and proteins. Conclusions: ML identified clinical phenotypes showed different degrees of sepsis severity and could be translated to the plasma proteome. Plasma proteomics provided novel insights into the molecular processes of sepsis and allowed the characterization of the phenotypes. This approach represents a blueprint to identify molecular features of sepsis subgroups and may pave the way for future targeted sepsis therapy.

Project description:Background The heterogeneity of sepsis represents a major problem for the development of personalized sepsis therapies. Thus, sepsis subtyping emerged as an important tool to approach this problem, but little progress was made due to insufficient molecular insights. Modern proteomics techniques allow the identification of subtypes and enable molecular and mechanistical insights. Here, we analyzed a prospective multi-center sepsis cohort using plasma proteomics to describe and characterize sepsis plasma proteome subtypes. Methods Plasma samples from 333 patients collected at days 1 and 4 of sepsis were analyzed using liquid-chromatography coupled to tandem mass spectrometry. Plasma proteome subtypes were identified using K-means clustering and were characterized based on clinical routine data, cytokine measurements and proteomics data. A random forest machine learning (ML) classifier was generated to enable the assignment of patients to the subtypes in future. Results Four subtypes with different sepsis severity were identified. Cluster 0 represented the most severe sepsis with 100 % mortality. Cluster 1, 2 and 3 showed a gradual decrease of the median SOFA score, which was reflected by clinical data and cytokine measurements. On the proteome level, the subtypes were characterized by distinct molecular features. We found an alternating immune response with cluster 1 showing prominent activation of the adaptive immune system as indicated by elevated levels of immunoglobulins (Ig) that were verified using orthogonal measurements. Cluster 2 was characterized by acute inflammation and the lowest Ig levels. Cluster 3 represented the sepsis proteome baseline of the investigated cohort. We generated a ML classifier and optimized it for a minimum number of proteins that could realistically be implemented into routine diagnostics. The final model was based on 10 proteins and Ig quantities and allowed the assignment of patients to clusters 1, 2 and 3 with high confidence. Conclusion The identified plasma proteome subtypes provide insights into immune response and disease mechanisms and allow conclusions on appropriate therapeutic measures. Thus, they represent a step forward in the development of targeted therapies and personalized medicine in sepsis.

Project description:Two molecular phenotypes of sepsis and acute respiratory distress syndrome, termed hyperinflammatory and hypoinflammatory, have been consistently identified by latent class analysis in numerous cohorts, with widely divergent clinical outcomes and differential responses to some treatments; however, the key biological differences between these phenotypes remain poorly understood. We used host and microbe metagenomic sequencing data from blood to deepen our understanding of biological differences between latent class analysis-derived phenotypes and to assess concordance between the latent class analysis-derived phenotypes and phenotypes reported by other investigative groups (e.g., SRS1-2, MARS1-4, reactive/uninflamed). We analyzed data from 113 hypoinflammatory and 76 hyperinflammatory sepsis patients enrolled in a two-hospital prospective cohort study. Molecular phenotypes had been previously assigned using latent class analysis. The hyperinflammatory and hypoinflammatory phenotypes of sepsis had distinct gene expression signatures, with 5,755 genes (31%) differentially expressed. The hyperinflammatory phenotype was associated with elevated expression of innate immune response genes, while the hypoinflammatory phenotype was associated with elevated expression of adaptive immune response genes, and notably, T-cell response genes. Plasma metagenomic analysis identified differences in prevalence of bacteremia, bacterial DNA abundance and composition between the phenotypes, with an increased presence and abundance of Enterobacteriaceae in the hyperinflammatory phenotype. Significant overlap was observed between these phenotypes and previously identified transcriptional subtypes of acute respiratory distress syndrome (reactive/uninflamed) and sepsis (SRS1-2). Analysis of data from the VANISH trial indicated that corticosteroids might have a detrimental effect in hypoinflammatory patients. The hyperinflammatory and hypoinflammatory phenotypes have distinct transcriptional and metagenomic features that could be leveraged for precision treatment strategies.

Project description:The wide array of molecules carried by plasma regulates critical immune functions and constitutes valuable biomarkers and therapeutic targets. In recent years the introduction of “systems approaches” has provided investigators with powerful means for assessing immune responses in patient samples on a global scale. However, while the use of genome-wide profiling technologies has become widespread, measuring the plasma proteome still presents considerable challenges. An alternative approach that consists in measuring transcriptome responses in reporter cells exposed in vitro to patient plasma has been successfully employed in a limited number of studies. Here we devised such a “Transcriptomic Reporter Assay” system to assess the immunogenicity of plasma from septic patients and evaluate its potential for biomarker discovery. Sepsis is a common, severe systemic infectious process for which physicians still lack efficient diagnostic or prognostic tools. Of the three different cell reporter systems tested, neutrophils were identified as the most capable “plasma sensor”. Compared to peripheral blood mononuclear cells and dendritic cell preparations neutrophils were best able to discriminate between plasma from septic and control subjects and responded by upregulating a robust immune transcriptional program. Additionally, the amplitude of the neutrophil transcriptomic response was shown to be associated with disease severity in two additional sets of patients. Overall, our results demonstrate both the suitability and potential clinical relevance of a neutrophil reporter assay for assessing immunopathogenic processes in a complex and severe condition such as sepsis. Peripheral blood mononuclear cells (PBMCs) were isolated from two healthy donors. Plasma samples were obtained from patients with culture-confirmed sepsis (n=12) and from uninfected controls (n=12). PBMCs were cultured for 6 h in medium alone, plasma from patients with sepsis, plasma from uninfected controls, and LPS using a final concentration of 20%. Transcriptional profiles were acquired using Illumina HumanHT12 V4 BeadChips.

Project description:The wide array of molecules carried by plasma regulates critical immune functions and constitutes valuable biomarkers and therapeutic targets. In recent years the introduction of “systems approaches” has provided investigators with powerful means for assessing immune responses in patient samples on a global scale. However, while the use of genome-wide profiling technologies has become widespread, measuring the plasma proteome still presents considerable challenges. An alternative approach that consists in measuring transcriptome responses in reporter cells exposed in vitro to patient plasma has been successfully employed in a limited number of studies. Here we devised such a “Transcriptomic Reporter Assay” system to assess the immunogenicity of plasma from septic patients and evaluate its potential for biomarker discovery. Sepsis is a common, severe systemic infectious process for which physicians still lack efficient diagnostic or prognostic tools. Of the three different cell reporter systems tested, neutrophils were identified as the most capable “plasma sensor”. Compared to peripheral blood mononuclear cells and dendritic cell preparations neutrophils were best able to discriminate between plasma from septic and control subjects and responded by upregulating a robust immune transcriptional program. Additionally, the amplitude of the neutrophil transcriptomic response was shown to be associated with disease severity in two additional sets of patients. Overall, our results demonstrate both the suitability and potential clinical relevance of a neutrophil reporter assay for assessing immunopathogenic processes in a complex and severe condition such as sepsis. Polymorphonuclear neutrophils (PMNs) were isolated from two healthy donors. Plasma samples were obtained from patients with culture-confirmed sepsis (n=12) and from uninfected controls (n=12). PMNs were cultured for 6 hours in medium alone, plasma from patients with sepsis, plasma from uninfected controls, and LPS using a final concentration of 20%. Transcriptional profiles were acquired using Illumina HumanHT12 V4 BeadChips.

			Action	DRS
	C1_report.pg_matrix.tsv	Tabular
	C3_report.pg_matrix.tsv	Tabular
	EXI01942.raw	Raw
	EXI02003.raw	Raw
	EXI02017.raw	Raw