Related Experiment Video
Updated: Jun 14, 2025
09:43
Analyses of Proteinuria, Renal Infiltration of Leukocytes, and Renal Deposition of Proteins in Lupus-prone MRL/lpr Mice
Published on: June 8, 2022
2.8K
Applying 12 machine learning algorithms and Non-negative Matrix Factorization for robust prediction of lupus
Lisha Mou1,2, Ying Lu1,2, Zijing Wu1,2
1Department of Rheumatology and Immunology, Institute of Translational Medicine, Health Science Center, The First Affiliated Hospital of Shenzhen University, Shenzhen Second People's Hospital, Shenzhen, China.
Frontiers in Immunology
|September 3, 2024
Summary
Machine learning models identified key immune genes for diagnosing lupus nephritis (LN). Six hub genes show promise for noninvasive LN detection and understanding disease progression.
Area of Science:
- Immunology
- Genomics
- Computational Biology
Background:
- Lupus nephritis (LN) presents diagnostic and therapeutic challenges.
- Understanding immune cell roles in LN pathogenesis is crucial.
Purpose of the Study:
- To profile LN using machine learning on single-cell kidney biopsy data.
- To identify novel diagnostic markers and understand LN pathophysiology.
Main Methods:
- Applied 12 machine learning algorithms and Non-negative Matrix Factorization (NMF) to single-cell datasets.
- Developed 102 immune-related gene (IRG) predictive models.
- Validated models using Area Under the Curve (AUC) and external cohorts; performed protein-protein interaction (PPI) analysis.
Main Results:
- Identified key immune cell populations and their roles in LN.
- Highlighted six hub IRGs (CD14, CYBB, IFNGR1, IL1B, MSR1, PLAUR) as potent diagnostic markers.
- Demonstrated high diagnostic accuracy in renal and peripheral blood, correlating gene expression with GFR, proteinuria, and creatinine.
Conclusions:
- Novel machine learning-based IRG models offer a promising noninvasive approach for LN diagnosis.
- Hub IRGs like IFNGR1, PLAUR, and CYBB are implicated in LN pathophysiology and disease severity.
- Integration of genomic data and machine learning can advance personalized management of autoimmune diseases like LN.
