Extended Data Fig. 5: Performance of retrained CDR3β-only and CDR3β+others models on seen- and unseen-epitope predictions in terms of multiple metrics.
From: Assessment of computational methods in predicting TCR–epitope binding recognition
a-c, Performance of retrained CDR3β-only models in seen-epitope test (a), independent test (b) and unseen-epitope independent test (c) using AS negatives based on CDR3β-only data. d-e, Performance of retrained CDR3β+others models (d) and retrained CDR3β-only models (e) in seen-epitope test using AS negatives based on CDR3β+others data. f-g, Performance of retrained CDR3β+others models (f) and retrained CDR3β-only models (g) in seen-epitope independent test using AS negatives based on CDR3β+others data. h-i, Performance of retrained CDR3β+others models (h) and retrained CDR3β-only models (i) in unseen-epitope independent test using AS negatives based on CDR3β+others data. All heatmaps show results of multiple metrics: AUPRC, Precision, Specificity, Recall and F1.
