Prediction of hub genes in pulpal inflammation and regeneration using autoencoders and a generative AI approach.
Journal:
Scientific reports
Published Date:
Jul 19, 2025
Abstract
Pulpal inflammation and regeneration are crucial for enhancing endodontic treatment outcomes. Transcriptomic studies highlight the involvement of proinflammatory cytokines, NF-κB signaling, and stem cell activity. This study employs a generative AI approach to predict and reconstruct hub genes associated with these processes, providing insights into biological mechanisms and potential therapeutic targets. Differential gene expression analysis was performed on data from the accession number GSE255672 using the GEO2R tool and Cytoscape, a bioinformatics software platform. A protein-protein interaction network was constructed using gene ontology annotations to identify key genes and subnetworks. CytoHubba, a Cytoscape plugin, was used to pinpoint hub genes using the Maximal Clique Centrality method. The Dataset was normalized, cleaned, and categorized into hub and non-hub genes. The data was then split into 80% training and 20% test sets for analysis using autoencoders. Autoencoders, which reduce complex data into simplified feature sets, were employed to compress the data for classifier training. An autoencoder-based model was trained using the preprocessed dataset, demonstrating moderate predictive performance with an accuracy of 76.92%, a precision-recall AUC of 0.9214, and a ROC AUC of 0.7333. The model performed well, achieving good predictive accuracy. The autoencoder achieved an accuracy rate of 76.92%, indicating a balanced performance between precision and recall. The model exhibited strong performance in identifying positive cases, with an area under the precision-recall curve of 0.9214. While the model demonstrated a moderate correlation between predicted and actual classifications, there remains room for further optimization. This study demonstrates the potential utility of autoencoders in predicting hub genes involved in pulpal inflammation and regeneration. These findings aim to support personalized strategies for improving pulpal health.
