Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms

BackgroundThe bile acid metabolism (BAM) and fatty acid metabolism (FAM) have been implicated in Kawasaki disease (KD), but their precise mechanisms remain unclear. Identifying signature cells and genes related to BAM and FAM could offer a deeper understanding of their role in the pathogenesis of KD...

Full description

Saved in:
Bibliographic Details
Main Authors: Chenhui Feng, Zhimiao Wei, Xiaohui Li
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-04-01
Series:Frontiers in Immunology
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fimmu.2025.1541939/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849731007421349888
author Chenhui Feng
Zhimiao Wei
Xiaohui Li
Xiaohui Li
author_facet Chenhui Feng
Zhimiao Wei
Xiaohui Li
Xiaohui Li
author_sort Chenhui Feng
collection DOAJ
description BackgroundThe bile acid metabolism (BAM) and fatty acid metabolism (FAM) have been implicated in Kawasaki disease (KD), but their precise mechanisms remain unclear. Identifying signature cells and genes related to BAM and FAM could offer a deeper understanding of their role in the pathogenesis of KD.MethodWe analyzed the public single-cell RNA sequencing (scRNA-seq) dataset GSE1687323 to characterize the immune cell-type landscape in KD. Gene sets related to BAM and FAM were collected from the Gene Set Enrichment Analysis (GSEA) database and previous literature. We analyzed the cellular heterogeneity of BAM and FAM at the single-cell level using R packages. Through differential expressed genes (DEG) analysis, high-dimensional Weighted Correlation Network Analysis (hdWGCNA) and machine learning algorithms, we identified signature genes associated with both BAM and FAM. The cellular expression patterns of signature genes were further validated using our own scRNA-seq dataset. Finally, quantitative real-time PCR (qRT–PCR) was performed to validate the expression levels of signature genes in KD, and Receiver Operating Characteristic (ROC) curve analysis was conducted to evaluate their diagnostic potential.ResultsEnhanced BAM and FAM were detected in monocytes and natural killer (NK) cells from KD in the public scRNA-seq dataset. Our scRNA-seq data confirmed the signature genes identified by machine learning algorithms: Vimentin (VIM) and chloride intracellular channel 1 (CLIC1) were upregulated in monocytes, while integrin subunit beta 2 (ITGB2) was elevated in NK cells of KD. qRT-PCR results also validated the bioinformatic analysis. Moreover, these genes demonstrated significant diagnostic potential. In the training dataset (GSE68004), the area under the curve (AUC) values and 95% CI were as follows: VIM: 0.914 (0.863–0.966), ITGB2: 0.958 (0.925–0.991), and CLIC1: 0.985 (0.969–1). The validation dataset (GSE73461) yielded similarly robust results, with AUC values and 95% CI: VIM: 0.872 (0.811–0.934), ITGB2: 0.861 (0.795–0.928), and CLIC1: 0.893 (0.837–0.948).ConclusionThis study successfully identified and validated VIM and CLIC1 in monocytes, as well as ITGB2 in NK cells, as novel metabolism-related genes in KD. These findings suggest that BAM and FAM may play crucial roles in KD pathogenesis. Furthermore, these signature genes hold promising potential as diagnostic biomarkers for KD.
format Article
id doaj-art-e3439da089e64c93a432b18c9174f9eb
institution DOAJ
issn 1664-3224
language English
publishDate 2025-04-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Immunology
spelling doaj-art-e3439da089e64c93a432b18c9174f9eb2025-08-20T03:08:42ZengFrontiers Media S.A.Frontiers in Immunology1664-32242025-04-011610.3389/fimmu.2025.15419391541939Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithmsChenhui Feng0Zhimiao Wei1Xiaohui Li2Xiaohui Li3Capital Institute of Pediatrics-Peking University Teaching Hospital, Beijing, ChinaDepartment of Cardiovascular Medicine, Children’s Hospital Capital Institute of Pediatrics, Beijing, ChinaCapital Institute of Pediatrics-Peking University Teaching Hospital, Beijing, ChinaDepartment of Cardiovascular Medicine, Children’s Hospital Capital Institute of Pediatrics, Beijing, ChinaBackgroundThe bile acid metabolism (BAM) and fatty acid metabolism (FAM) have been implicated in Kawasaki disease (KD), but their precise mechanisms remain unclear. Identifying signature cells and genes related to BAM and FAM could offer a deeper understanding of their role in the pathogenesis of KD.MethodWe analyzed the public single-cell RNA sequencing (scRNA-seq) dataset GSE1687323 to characterize the immune cell-type landscape in KD. Gene sets related to BAM and FAM were collected from the Gene Set Enrichment Analysis (GSEA) database and previous literature. We analyzed the cellular heterogeneity of BAM and FAM at the single-cell level using R packages. Through differential expressed genes (DEG) analysis, high-dimensional Weighted Correlation Network Analysis (hdWGCNA) and machine learning algorithms, we identified signature genes associated with both BAM and FAM. The cellular expression patterns of signature genes were further validated using our own scRNA-seq dataset. Finally, quantitative real-time PCR (qRT–PCR) was performed to validate the expression levels of signature genes in KD, and Receiver Operating Characteristic (ROC) curve analysis was conducted to evaluate their diagnostic potential.ResultsEnhanced BAM and FAM were detected in monocytes and natural killer (NK) cells from KD in the public scRNA-seq dataset. Our scRNA-seq data confirmed the signature genes identified by machine learning algorithms: Vimentin (VIM) and chloride intracellular channel 1 (CLIC1) were upregulated in monocytes, while integrin subunit beta 2 (ITGB2) was elevated in NK cells of KD. qRT-PCR results also validated the bioinformatic analysis. Moreover, these genes demonstrated significant diagnostic potential. In the training dataset (GSE68004), the area under the curve (AUC) values and 95% CI were as follows: VIM: 0.914 (0.863–0.966), ITGB2: 0.958 (0.925–0.991), and CLIC1: 0.985 (0.969–1). The validation dataset (GSE73461) yielded similarly robust results, with AUC values and 95% CI: VIM: 0.872 (0.811–0.934), ITGB2: 0.861 (0.795–0.928), and CLIC1: 0.893 (0.837–0.948).ConclusionThis study successfully identified and validated VIM and CLIC1 in monocytes, as well as ITGB2 in NK cells, as novel metabolism-related genes in KD. These findings suggest that BAM and FAM may play crucial roles in KD pathogenesis. Furthermore, these signature genes hold promising potential as diagnostic biomarkers for KD.https://www.frontiersin.org/articles/10.3389/fimmu.2025.1541939/fullKawasaki diseasebile acid metabolismfatty acid metabolismsingle-cell RNA sequencingmachine learning
spellingShingle Chenhui Feng
Zhimiao Wei
Xiaohui Li
Xiaohui Li
Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms
Frontiers in Immunology
Kawasaki disease
bile acid metabolism
fatty acid metabolism
single-cell RNA sequencing
machine learning
title Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms
title_full Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms
title_fullStr Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms
title_full_unstemmed Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms
title_short Identification of novel metabolism-related biomarkers of Kawasaki disease by integrating single-cell RNA sequencing analysis and machine learning algorithms
title_sort identification of novel metabolism related biomarkers of kawasaki disease by integrating single cell rna sequencing analysis and machine learning algorithms
topic Kawasaki disease
bile acid metabolism
fatty acid metabolism
single-cell RNA sequencing
machine learning
url https://www.frontiersin.org/articles/10.3389/fimmu.2025.1541939/full
work_keys_str_mv AT chenhuifeng identificationofnovelmetabolismrelatedbiomarkersofkawasakidiseasebyintegratingsinglecellrnasequencinganalysisandmachinelearningalgorithms
AT zhimiaowei identificationofnovelmetabolismrelatedbiomarkersofkawasakidiseasebyintegratingsinglecellrnasequencinganalysisandmachinelearningalgorithms
AT xiaohuili identificationofnovelmetabolismrelatedbiomarkersofkawasakidiseasebyintegratingsinglecellrnasequencinganalysisandmachinelearningalgorithms
AT xiaohuili identificationofnovelmetabolismrelatedbiomarkersofkawasakidiseasebyintegratingsinglecellrnasequencinganalysisandmachinelearningalgorithms