Enhancing TCR specificity predictions by combined pan- and peptide-specific training, loss-scaling, and sequence similarity integration

通过结合泛特异性和肽特异性训练、损失缩放和序列相似性整合来增强TCR特异性预测。