Fig. 4.

Comparison of the TAA prediction accuracy of TAPINTO and other methods. This figure evaluates the sensitivity of various methods in identifying established TAAs using ROC/AUC analysis. (A) Sensitivity for receptor tyrosine kinase (RTK) TAAs: Comparison of prediction accuracy using TAPINTO-calculated fold change (red line) versus traditional differential expression methods, including LIMMA (blue line) and t test (green line), based on well-known breast cancer targets. (B) Comparison with existing outlier detection: Performance benchmarking of TAPINTO against DIDS (deep blue line), a conventional outlier-based detection method, specifically evaluating the ability to identify oncogenic RTK targets within heterogeneous populations. (C and D) Generalizability across diverse antigen families: Evaluation of prediction sensitivity for various TAA categories, including CEA-, MMP-, FZD-, and MUC-encoding genes. TAPINTO’s performance is compared against (C) mean-based statistical tests (LIMMA and t test) and (D) the specific fold change logic of DIDS.