Dataset Information

Pediatric Crohn's disease diagnosis aid via genomic analysis and machine learning.

ABSTRACT:

Introduction

Determination of pediatric Crohn's disease (CD) remains a major diagnostic challenge. However, the rapidly emerging field of artificial intelligence has demonstrated promise in developing diagnostic models for intractable diseases.

Methods

We propose an artificial neural network model of 8 gene markers identified by 4 classification algorithms based on Gene Expression Omnibus database for diagnostic of pediatric CD.

Results

The model achieved over 85% accuracy and area under ROC curve value in both training set and testing set for diagnosing pediatric CD. Additionally, immune infiltration analysis was performed to address why these markers can be integrated to develop a diagnostic model.

Conclusion

This study supports further clinical facilitation of precise disease diagnosis by integrating genomics and machine learning algorithms in open-access database.

SUBMITTER: Zheng Z

PROVIDER: S-EPMC10076664 | biostudies-literature | 2023

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Pediatric Crohn's disease diagnosis aid <i>via</i> genomic analysis and machine learning.

Zheng Zhiwei Z Zhan Sha S Zhou Yongmao Y Huang Ganghua G Chen Pan P Li Baofei B

Frontiers in pediatrics 20230323

<h4>Introduction</h4>Determination of pediatric Crohn's disease (CD) remains a major diagnostic challenge. However, the rapidly emerging field of artificial intelligence has demonstrated promise in developing diagnostic models for intractable diseases.<h4>Methods</h4>We propose an artificial neural network model of 8 gene markers identified by 4 classification algorithms based on Gene Expression Omnibus database for diagnostic of pediatric CD.<h4>Results</h4>The model achieved over 85% accuracy ...[more]

PMID: 37033178

Similar Datasets

Project description:BackgroundThe incidence of pediatric Crohn's disease (PCD) is increasing worldwide every year. The challenges in early diagnosis and treatment of PCD persist due to its inherent heterogeneity. This study's objective was to discover novel diagnostic markers and molecular subtypes aimed at enhancing the prognosis for patients suffering from PCD.MethodsCandidate genes were obtained from the GSE117993 dataset and the GSE93624 dataset by weighted gene co-expression network analysis (WGCNA) and differential analysis, followed by intersection with platelet-related genes. Based on this, diagnostic markers were screened by five machine learning algorithms. We constructed predictive models and molecular subtypes based on key markers. The models were evaluated using the GSE101794 dataset as the validation set, combined with receiver operating characteristic curves, decision curve analysis, clinical impact curves, and calibration curves. In addition, we performed pathway enrichment analysis and immune infiltration analysis for different molecular subtypes to assess their differences.ResultsThrough WGCNA and differential analysis, we successfully identified 44 candidate genes. Following this, employing five machine learning algorithms, we ultimately narrowed it down to five pivotal markers: GNA15, PIK3R3, PLEK, SERPINE1, and STAT1. Using these five key markers as a foundation, we developed a nomogram exhibiting exceptional performance. Furthermore, we distinguished two platelet-related subtypes of PCD through consensus clustering analysis. Subsequent analyses involving pathway enrichment and immune infiltration unveiled notable disparities in gene expression patterns, enrichment pathways, and immune infiltration landscapes between these subtypes.ConclusionIn this study, we have successfully identified five promising diagnostic markers and developed a robust nomogram with high predictive efficacy. Furthermore, the recognition of distinct PCD subtypes enhances our comprehension of potential pathogenic mechanisms and paves the way for future prospects in early diagnosis and personalized treatment.