Related Experiment Video
Updated: Oct 28, 2026
Closed-Loop Neurostimulation for Biomarker-Driven, Personalized Treatment of Major Depressive Disorder
Published on: July 7, 2023
The Construction and Validation of a Machine Learning Model for Major Depressive Disorder Based on Astrocytes and
Shi Lin1, Gaoyuan Wang1, Xiaoqin Shao1
1Research Centre of Basic Integrative Medicine, School of Basic Medical Science, Guangzhou University of Chinese Medicine, Guangzhou 510006, China.
Background:
Astrocytes play an important role in depression. However, their specific mechanisms, especially their subcellular mechanisms, remain unclear.
Objective:
We aimed to identify key genes associated with mitochondrial dynamics and astrocytes in Major Depressive Disorder (MDD), construct a diagnostic prediction model, and elucidate their regulatory mechanisms in neuroinflammation.
Methods:
Depression-related transcriptome datasets were retrieved from the GEO database. Overlapping differentially expressed genes between mitochondrial dynamics-related and astrocytes-associated genes were screened, followed by functional enrichment analysis. Machine learning algorithms were employed to establish a diagnostic model and identify hub genes. An LPS-induced mouse depression model was established, and depressive-like phenotypes were assessed through behavioral tests. Hub gene expression was validated by using quantitative real-time PCR and Western blotting.
Results:
A machine learning model with favorable diagnostic performance was constructed. LPS markedly upregulated hexokinase 2 (HK2) and ribosomal protein S7 (RPS7) expression in the prefrontal cortex and hippocampus, suggesting that glycolytic metabolism and ribosomal function contribute to LPS-induced neuropathological changes.
Conclusions:
HK2 and RPS7 were identified and validated as crucial hub genes responsive to neuroinflammation. The combined action of mitochondrial dynamics and astrocyte metabolic abnormalities may play a key role in depression's pathogenesis, providing novel candidate biomarkers and theoretical support for the molecular diagnosis of MDD and anti-inflammatory targeted therapy.
