Identification and validation of core genes associated with programmed cell death in the early stage of severe trauma using machine learning and neural networks
Yao Le
Liu Zhibing
Han Zhenyuan
Lu Guanyu
Wang Xin
Abstract：Objective To identify and validate core genes related to programmed cell death(PCD)for ear-ly severe trauma(≤12 h post-injury)and to explore their biological significance.Methods Gene expression data-sets from the GEO database were standardized and split into training/testing sets.Differential analysis identified 19 PCD-related genes.Protein-protein interaction network analysis revealed hub genes,further refined by five machine learning algorithms(random forest,least absolute shrinkage and selection operator,XGBoost,gradient boosting ma-chine,elastic net).Core genes were validated via ROC curves and neural networks.Results From the 416 PCD-related differential expressed genes,10 hub genes emerged.Four core genes(STAT3,IL-10,HDAC1,PIK3R1)showed consistent differentially expression and high predictive accuracy(AUC＞0.9)across the datasets.Conclu-sion A stepwise machine learning approach was employed to identify core PCD-related genes in the early stages of severe trauma.This study investigated the potential mechanisms by which these genes modulate immune homeostasis remodeling and inflammatory responses through regulatory pathways of programmed cell death,thereby contributing to trauma-induced pathological progression.However,the specific molecular interactions underlying these observations warrant further experimental validation.
Keywords：Severe traumaProgrammed cell deathMachine learning
Publication Date：2025-05-15
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：5( 389-393 )