Machine learning-based identification of cuproptosis-related markers and immune infiltration in severe community-acquired pneumonia

Clin Respir J. 2023 Jul;17(7):618-628. doi: 10.1111/crj.13633. Epub 2023 Jun 6.

Abstract

Background: Severe community-acquired pneumonia (SCAP) is one of the world's most common diseases and a major etiology of acute respiratory distress syndrome (ARDS). Cuproptosis is a novel form of regulated cell death that can occur in various diseases.

Methods: Our study explored the degree of immune cell infiltration during the onset of severe CAP and identified potential biomarkers related to cuproptosis. Gene expression matrix was obtained from GEO database indexed GSE196399. Three machine learning algorithms were applied: The least absolute shrinkage and selection operator (LASSO), the random forest, and the support vector machine-recursive feature elimination (SVM-RFE). Immune cell infiltration was quantified by single-sample gene set enrichment analysis (ssGSEA) scoring. Nomogram was constructed to verify the applicability of using cuproptosis-related genes to predict the onset of severe CAP and its deterioration toward ARDS.

Results: Nine cuproptosis-related genes were differentially expressed between the severe CAP group and the control group: ATP7B, DBT, DLAT, DLD, FDX1, GCSH, LIAS, LIPT1, and SLC31A1. All 13 cuproptosis-related genes were involved in immune cell infiltration. A three-gene diagnostic model was constructed to predict the onset of severe CAP: GCSH, DLD, and LIPT1.

Conclusion: Our study confirmed the involvement of the newly discovered cuproptosis-related genes in the progression of SCAP.

Keywords: ARDS; bioinformatics; cuproptosis; severe CAP.

MeSH terms

  • Algorithms
  • Apoptosis*
  • Community-Acquired Infections*
  • Copper
  • Humans
  • Machine Learning
  • Pneumonia*
  • Respiratory Distress Syndrome*

Substances

  • Copper