Fig. 2: Results of UniPMT compared with the baselines on P–M–T, P–M and P–T binding prediction tasks. | Nature Machine Intelligence

Fig. 2: Results of UniPMT compared with the baselines on P–M–T, P–M and P–T binding prediction tasks.

From: A unified deep framework for peptide–major histocompatibility complex–T cell receptor binding prediction

Fig. 2

a,b, ROC-AUC (a) and PR-AUC (b) of UniPMT and the baselines on the P–M–T dataset. c,d, ROC-AUC (c) and PR-AUC (d) of UniPMT and the baselines on the P–M dataset (shown as mean values ± s.d.). e,f, ROC-AUC (e) and PR-AUC (f) of UniPMT and the baselines on the P–M dataset and among the seven alleles with more than 1,000 pairs in the testing set. g,h, ROC-AUC (g) and PR-AUC (h) of UniPMT and the baselines on the P–T TEIM dataset. i,j, ROC-AUC (i) and PR-AUC (j) of UniPMT and the baselines on the P–T zero-shot dataset.

Source data

Back to article page