Predictive Modelling of PANoptosis related Biomarkers in Asthma based on convolutional neural network algorithm

BACKGROUND Bronchial asthma is a chronic inflammatory airway disease that ranks 3rd among respiratory diseases, but there is still no curative treatment. PANoptosis has been shown to play an important role in a variety of inflammatory diseases, but its relevance in asthma remains poorly studied. OBJECTIVE To investigate PANoptosis-related characteristic genes and their associated functional features in asthma. METHODS Gene information was obtained mainly from the GEO database and GeneCards database. Gene function identification was based on GO database and KEGG database. Gene set enrichment analysis was used to identify signalling pathways enriched for PANoptosis related genes in asthma. We made three machine learning models, support vector machine, LASSO regression analysis and random forest regression, to identify characteristic PANoptosis-related genes in asthma. The lncRNA-miRNA-mRNA network was constructed using the STRING gene database search tool and Cytoscape. The Pearson algorithm was used to analyse immune cell infiltration and to analyse the correlation between immune cell abundance and PANoptosis-related differentially expressed genes. ROC curves were used to analyse the accuracy of potential target gene diagnosis. Predictive models of PANoptosis-related biomarkers in asthma were constructed by convolutional neural networks. RESULTS We finally obtained 10 asthma-associated characteristic PANoptosis-related genes, and their high-expression groups were mainly enriched in the entries of Leishmania protozoa infection, cytokine-receptor interactions, NOD-like receptor signalling pathway, lysosome, and Toll-like receptor signalling pathway, which are mostly pro-inflammatory and immunoregulatory pathways between immune cells. The level of M0 macrophage infiltration was higher in asthma group than in the normal group. Finally, we also confirmed that the predictive model of PANoptosis-related biomarkers has a good predictive efficacy in asthma. CONCLUSION We demonstrated for the first time the role of PANoptosishub genes in asthma, which provides new ideas for further research on the mechanism of asthma development and therapeutic strategies.