Related Experiment Video
Updated: Jun 23, 2025
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.8K
A Machine Learning Framework for Screening Plasma Cell-Associated Feature Genes to Estimate Osteoporosis Risk and
Shoubao Wang1, Jiafu Zhu2, Weinan Liu3,4
1Department of Orthopedics, Huai'an Hospital Affiliated to Yangzhou University (The Fifth People's Hospital of Huai'an), Huai'an, 223300, China.
Biochemical Genetics
|June 19, 2024
Summary
This study uses machine learning to identify genes in circulating monocytes linked to osteoporosis. These findings aid in personalized risk assessment and treatment strategies for bone fragility.
Area of Science:
- Genomics
- Computational Biology
- Bone Biology
Background:
- Osteoporosis is a major global health issue characterized by fragile bones due to low bone density.
- Bone mineral density (BMD) is crucial for diagnosing osteoporosis, and circulating monocytes are key players in bone remodeling.
- Identifying molecular markers associated with bone loss can improve osteoporosis management.
Purpose of the Study:
- To develop a machine learning framework to identify circulating monocyte-associated genes impacting bone loss in osteoporosis.
- To explore the diagnostic and prognostic potential of these genes for osteoporosis.
- To uncover potential therapeutic targets and understand the molecular mechanisms involved in osteoporosis.
Main Methods:
- Utilized publicly available datasets (GSE56815, GSE7158, GSE7429, GSE62402) from female patients with varying BMD.
- Quantified monocyte types using CIBERSORT and identified differentially expressed genes (DEGs).
- Employed machine learning models (GLM, Random Forest, XGB, SVM) for feature selection, followed by Artificial Neural Networks (ANNs) and nomograms for diagnosis and risk estimation.
Main Results:
- Support Vector Machines (SVM) demonstrated superior performance in identifying key genes (DEFA4, HLA-DPB1, LCN2, HP, GAS7) associated with osteoporosis.
- ANNs and nomograms effectively distinguished between low and high BMD, estimating osteoporosis risk.
- Identified potential therapeutic agents (e.g., clozapine, aspirin) and highlighted the role of miRNAs, m6A modifications, and autophagy in regulating identified genes.
Conclusions:
- A machine learning framework utilizing monocyte-associated genes offers a promising approach for personalized risk stratification in osteoporosis.
- The identified genes and pathways provide insights into the molecular mechanisms of osteoporosis and potential therapeutic interventions.
- This approach can aid in estimating treatment vulnerability and guiding personalized medicine for osteoporosis patients.
