Related Experiment Video
Updated: Jun 2, 2025
07:15
Robust Ligature-Induced Model of Murine Periodontitis for the Evaluation of Oral Neutrophils
Published on: January 21, 2020
11.3K
Variational graph autoencoder for reconstructed transcriptomic data associated with NLRP3 mediated pyroptosis in
Pradeep K Yadalam1, Prabhu Manickam Natarajan2, Carlos M Ardila3
1Department of Periodontics, Saveetha Dental College, Saveetha Institute of Medical and Technology Sciences, SIMATS, Saveetha University, Chennai, 600077, Tamil Nadu, India.
Scientific Reports
|January 14, 2025
Summary
This study shows Variational Graph Autoencoders (VGAEs) can accurately reconstruct gene data for periodontitis research. These deep learning models show promise for analyzing gene expression and understanding inflammatory processes like pyroptosis.
Area of Science:
- Oral biology
- Immunology
- Bioinformatics
Background:
- Periodontitis involves Porphyromonas gingivalis, leading to inflammation and bone loss.
- The NLRP3 inflammasome and pyroptosis are key inflammatory pathways in periodontal disease.
- Hypoxia in periodontal disease exacerbates inflammation and bacterial survival.
Purpose of the Study:
- To evaluate Variational Graph Autoencoders (VGAEs) for reconstructing gene expression data in periodontitis.
- To analyze gene data related to NLRP3 inflammasome-mediated pyroptosis under hypoxic conditions.
- To assess the performance of VGAE and K-means clustering in periodontitis gene expression analysis.
Main Methods:
- Utilized NCBI GEO dataset GSE262663 with samples exposed to hypoxia.
- Applied unsupervised K-means clustering for identifying natural groupings in biological data.
- Employed Variational Graph Autoencoders (VGAEs), a deep learning model, for graph structure reconstruction and gene data analysis.
Main Results:
- VGAE achieved high performance with 99.42% accuracy and perfect precision.
- The model accurately predicted 4,080 out of 9,900 positive samples, with 5,820 false negatives.
- VGAE's latent space showed significant differences from original data, indicating clustered gene expression patterns.
Conclusions:
- VGAE demonstrates significant potential for accurate gene expression pattern reconstruction in periodontitis research.
- K-means clustering and VGAE show promise as tools for analyzing complex biological data in periodontitis.
- These methods can aid in understanding gene expression dynamics and inflammatory pathways in periodontal disease.
