Immune status assessment based on plasma proteomics with meta graph convolutional networks.

Area of Science:

• Biochemistry
• Immunology
• Computational Biology

Background:

• Plasma proteins are crucial for immune health assessment and disease risk prediction.
• The relationship between plasma proteins and systemic immune function requires further elucidation.
• Developing advanced computational models is essential for analyzing complex proteomic data.

Purpose of the Study:

• To develop and validate a novel computational framework, ProMetaGCN, for evaluating systemic immune status using plasma proteomics.
• To identify key immune-related proteins and their associated biological pathways.
• To establish a link between immune profiles, disease progression (COVID-19), aging, and patient survival (NSCLC).

Main Methods:

• Integration of meta-learning, graph convolutional networks, and protein-protein interaction (PPI) data.
• Application of machine learning algorithms (Random Forest, LightGBM, XGBoost, Lasso) for immune profiling and biomarker discovery.
• Validation of the model using independent COVID-19 patient cohorts and analysis of non-small-cell lung cancer (NSCLC) patient survival data.

Main Results:

• Identification of 309 immune-related factors and their biological functions/pathways.
• Discovery of ADAMTS13, GDF15, and SERPINF2 as significant biomarkers for immune profiling and aging.
• Demonstration of correlation between immune status and COVID-19 infection progression/recovery.
• Introduction of ImmuneAgeGap metric showing association with survival rates in NSCLC patients.

Conclusions:

• ProMetaGCN provides a robust framework for assessing immune status through plasma proteomics.
• Key identified biomarkers and the ImmuneAgeGap metric offer potential for personalized medicine and disease risk stratification.
• The study advances understanding of immune function and its implications in infectious diseases and cancer.