Machine learning algorithm predicts fibrosis-related blood diagnosis markers of intervertebral disc degeneration

Abstract

Abstract Background Intervertebral disc cell fibrosis has been established as a contributing factor to intervertebral disc degeneration (IDD). This study aimed to identify fibrosis-related diagnostic genes for patients with IDD. Methods RNA-sequencing data was downloaded from Gene Expression Omnibus (GEO) database. The diagnostic genes was identified using Random forest based on the differentially expressed fibrosis-related genes (DE-FIGs) between IDD and control samples. The immune infiltration states in IDD and the regulatory network as well as potential drugs targeted diagnostic genes were investigated. Quantitative Real-Time PCR was conducted for gene expression valifation. Results CEP120 and SPDL1 merged as diagnostic genes. Substantial variations were observed in the proportions of natural killer cells, neutrophils, and myeloid-derived suppressor cells between IDD and control samples. Further experiments indicated that AC144548.1 could regulate the expressions of SPDL1 and CEP120 by combininghsa-miR-5195-3p and hsa-miR-455-3p, respectively. Additionally, transcription factors FOXM1, PPARG, and ATF3 were identified as regulators of SPDL1 and CEP120 transcription. Notably, 56 drugs were predicted to target these genes. The down-regulation of SPDL1 and CEP120 was also validated. Conclusion This study identified two diagnostic genes associated with fibrosis in patients with IDD. Additionally, we elucidated their potential regulatory networks and identified target drugs, which offer a theoretical basis and reference for further study into fibrosis-related genes involved in IDD.