TrambaHLApan: a transformer and mamba-based neoantigen prediction method considering both antigen presentation and immunogenicity

Neoantigens represent ideal targets for tumor immunotherapy. This study introduces TrambaHLApan, a neoantigen prediction method that utilizes Transformer and Mamba models to consider both antigen presentation potential (TrambaHLApan-EL) and immunogenicity (TrambaHLApan-IM).

Data S1: Antigen presentation training set
Data S2: Immunogenicity training set
Data S3: Allele24 independent antigen presentation test dataset
Data S4: Allele36 independent antigen presentation test set
Data S5: Cancer neoantigen immunogenicity test set (GBM)
Data S6: Cancer neoantigen immunogenicity test set (MANAFEST)
fivefold_val_flags(DataS1): Five-fold cross-validation division file for DataS1
fivefold_val_flags(DataS2): Five-fold cross-validation division file for DataS2

新抗原（neoantigens）是肿瘤免疫治疗的理想靶点。本研究提出了TrambaHLApan这一新抗原预测方法，该方法结合Transformer与Mamba模型，同时考量抗原呈递潜能（TrambaHLApan-EL）与免疫原性（TrambaHLApan-IM）。
数据S1：抗原呈递训练集
数据S2：免疫原性训练集
数据S3：24等位基因独立抗原呈递测试数据集
数据S4：36等位基因独立抗原呈递测试集
数据S5：癌症新抗原免疫原性测试集（GBM）
数据S6：癌症新抗原免疫原性测试集（MANAFEST）
fivefold_val_flags(DataS1)：DataS1的五折交叉验证划分文件
fivefold_val_flags(DataS2)：DataS2的五折交叉验证划分文件