Skip to main content
Fig. 3 | BMC Bioinformatics

Fig. 3

From: MATHLA: a robust framework for HLA-peptide binding prediction integrating bidirectional LSTM and multiple head attention mechanism

Fig. 3

a AUC scores of three pan-allele models (MATHLA, netMHCpan 4.0 and ACME) over 10 non-overlapping alleles between the training and test datasets. b–d The receiver operating characteristic curve of MATHLA, netMHCpan 4.0 and MHCflurry on the test dataset. Plots for HLA-A, HLA-B and HLA-C alleles are generated separately. e The AUC scores of MATHLA and netMHCpan 4.0 over 21 HLA-C alleles in the test dataset

Back to article page