AUTHOR=Guo Yangyang , Cen Kenan , Hong Kai , Mai Yifeng , Jiang Minghui 

TITLE=Construction of a neural network diagnostic model for renal fibrosis and investigation of immune infiltration characteristics

JOURNAL=Frontiers in Immunology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/immunology/articles/10.3389/fimmu.2023.1183088

DOI=10.3389/fimmu.2023.1183088

ISSN=1664-3224

ABSTRACT=Background: Recently, the incidence rate of renal fibrosis is increasing worldwide, which greatly increased the burden to society. However, the diagnostic and therapeutic tools available for the disease are insufficient, which manifests the urgency to screen the potential biomarkers to predict renal fibrosis.
Methods: Based on the Gene Expression Omnibus (GEO) database, we obtained two  gene array datasets (GSE76882 and GSE22459) from patients with renal fibrosis and healthy people. We identified differentially expressed genes (DEGs) between renal fibrosis and normal tissues, and analyzed possible diagnostic biomarkers through  machine learning. The diagnostic effect of candidate markers was evaluated by receiver operating characteristic (ROC) curve. Reverse transcription quantitative polymerase chain reaction (RT-qPCR) was used to detect the expression of candidate biomarkers. Moreover, CIBERSORT algorithm was carried out to determine the proportion of 22 kinds of immune cells in patients with renal fibrosis, and the correlation between the expression of biomarkers and the proportion of immune cells was studied. Finally, we built an artificial neural network model for renal fibrosis.
Results: Foue candidate genes were identified as biomarkers of renal fibrosis, namely DOCK2, SLC1A3, SOX9 and TARP, with AUC values all higher than 0.75. After that, we verified the expression of these genes by RT-qPCR. Subsequently, we revealed the potential disorder immune cells in renal fibrosis group through CIBERSORT analysis, and found the immune cells were highly correlated related to the expression of candidate markers.
Conclusion: DOCK2, SLC1A3, SOX9 and TARP were identified as potential diagnostic genes for renal fibrosis, and their most relevant immune cells were revealed. The findings in our research provide possible biomarkers to diagnose renal fibrosis.