Machine learning-based characterization of cuprotosis-related biomarkers and immune infiltration in Parkinson’s disease

• DOI: 10.3389/fgene.2022.1010361
• Publication Date: 2022-10-17T08:57:43Z
• Authors: Songyun Zhao; Li Zhang; Wei Ji; Yachen Shi; Guichuan Lai; Hao Chi; Weiyi Huang; Chao Cheng
• Keywords: Nomogram

Background: Parkinson’s disease (PD) is a neurodegenerative commonly seen in the elderly. On other hand, cuprotosis new copper-dependent type of cell death that can be observed various diseases. Methods: This study aimed to identify potential novel biomarkers by biomarker analysis and explore immune infiltration during onset cuprotosis. Gene expression profiles were retrieved from GEO database for GSE8397, GSE7621, GSE20163, GSE20186 datasets. Three machine learning algorithms: least absolute shrinkage selection operator (LASSO), random forest, support vector machine-recursive feature elimination (SVM-RFE) used screen signature genes cuprotosis-related (CRG). Immune was estimated ssGSEA, associated with cells function examined using spearman correlation analysis. Nomogram created validate accuracy these predicting PD progression. Classification specimens consensus clustering methods. Result: datasets Expression Omnibus (GEO) combined after eliminating batch effects. By we identified three ATP7A, SLC31A1, DBT or more accurate diagnosis course. Patients could benefit clinically characteristic line graph based on genes. Consistent two subtypes, C2 subtype exhibiting higher function. Conclusion: In conclusion, our reveals several newly intervene progression through infiltration.