학술논문

Proteomic analysis to identification of hypoxia related markers in spinal tuberculosis: a study based on weighted gene co-expression network analysis and machine learning
Document Type
article
Source
BMC Medical Genomics, Vol 16, Iss 1, Pp 1-17 (2023)
Subject
Spinal tuberculosis
WGCNA
Machine learning
ssGSEA
Pharmaco-transcriptomic analysis
Internal medicine
RC31-1245
Genetics
QH426-470
Language
English
ISSN
1755-8794
Abstract
Abstract Objective This article aims at exploring the role of hypoxia-related genes and immune cells in spinal tuberculosis and tuberculosis involving other organs. Methods In this study, label-free quantitative proteomics analysis was performed on the intervertebral discs (fibrous cartilaginous tissues) obtained from five spinal tuberculosis (TB) patients. Key proteins associated with hypoxia were identified using molecular complex detection (MCODE), weighted gene co-expression network analysis(WGCNA), least absolute shrinkage and selection operator (LASSO), and support vector machine recursive feature Elimination (SVM-REF) methods, and their diagnostic and predictive values were assessed. Immune cell correlation analysis was then performed using the Single Sample Gene Set Enrichment Analysis (ssGSEA) method. In addition, a pharmaco-transcriptomic analysis was also performed to identify targets for treatment. Results The three genes, namely proteasome 20 S subunit beta 9 (PSMB9), signal transducer and activator of transcription 1 (STAT1), and transporter 1 (TAP1), were identified in the present study. The expression of these genes was found to be particularly high in patients with spinal TB and other extrapulmonary TB, as well as in TB and multidrug-resistant TB (p-value