Cell type predictions recapitulate known effects of aging on the immune system. (A) Overview of the study design, including the development of cell type prediction and age prediction models using single-cell transcriptomics. The workflow highlights data preprocessing, clustering, annotation, and model training. (B) Cell type proportion changes in individuals from different age groups. (C) CD4/CD8 ratio increases with age, normalized to individuals aged 18–35. (D) Comparison of predicted versus manually annotated cell types in the training dataset. Cell annotations were based on canonical markers: CD4, CD8A, CCR7, GZMB, GNLY, and FOXP3. Predicted clusters align closely with ground truth annotations, demonstrating the accuracy of the model. (E) Validation of the model in the test dataset, showing high concordance between predicted and manually annotated clusters with quantitative accuracy metrics (97% accuracy; F1 score = 0.97). (F) External validation using the Yasumizu et al. (2024) dataset demonstrates robustness across datasets (83% accuracy; F1 score = 0.80). Statistical significance is indicated: *** Bonferroni-corrected p-value less than or equal to .001, * Bonferroni-corrected P-value less than or equal to .05, # Bonferroni-corrected P-value less than or equal to .1.