Press "Enter" to skip to content



Please cite us if you find this resource useful:

[1]. Wu J, Zhao W, Zhou B, Su Z, Gu X, Zhou Z*, Chen S*. TSNAdb: a database for tumor-specific neoantigens from immunogenomics data analysis. Genomics  Proteomics Bioinformatics. 2018, 16(4),276–282. DOI: 10.1016/j.gpb.2018.06.003  pdf download

[2]. Zhou Z#, Lyu X#, Wu J, Yang X, Wu S, Zhou J, Gu X, Su Z*, Chen S*. TSNAD: an integrated software for cancer somatic mutation and tumour-specific neoantigen detection. R Soc Open Sci. 2017, 4: 170050. DOI: 10.1098/rsos.170050   pdf download

Data source

The somatic mutation data is extracted from The Cancer Genome Atlas(TCGA) and International Cancer Genome Consortium (ICGC);

The HLA allele information is extracted from The Cancer Immunome Atlas(TCIA)

The peptide-HLA binding validation data is extracted from The Immune Epitope Database(IEDB)

Prediction methods

          The binding affinities between peptides and HLA alleles are predicted by NetMHCpan 2.8/NetMHCpan 4.0.

The extensive analyses of potential neoantigens generated by somatic mutations of each gene are conducted by the tools embedded in TSNAD.

Data availability

Users can download data from the Download page. Data is freely available for academic users who agree with the TCGA publication guidelines .

Search guide

Since the results are predicted by two different versions of NetMHCpan (v2.8 and v4.0), users need to select  the preferred one (default: NetMHCpan 2.8). Then, users can input the gene name (e.g. KRAS)  to search the corresponding neoantigen results. The results will return in a few seconds.

In the ‘Browse‘ page, the gene name is required and the tissue name is optional in each retrieval.

In the result table, the description of each column is listed as follows:

         Tissue: tumor type.

         Mutation: amino acid change caused by the mutation. Users can obtain the detailed information by clicking on the mutation.

         Position in peptide: the mutation position in peptide (e.g. if the value is 8 and the peptide is EVFHACINWV, the mutant amino acid residue would be N).

         HLA allele: HLA allele.

         WT peptide: the sequence of wild-type peptide.

         WT affinity: the predicted binding affinity between wild-type peptide and HLA allele by NetMHCpan2.8 (4.0). 

         WT binding level: binding level between wild-type peptide and HLA allele.  ‘SB’ indicates strong binding (IC50 < 150nM) ,  ‘WB’ indicates weak binding (150nM < IC50  < 500nM),  ‘- ‘ indicates non-binding (IC50 > 500nM).

         MT peptide: the sequence of mutant peptide.

         MT affinity:the predicted binding affinity between mutant peptide and HLA allele by NetMHCpan2.8 (4.0).

         MT binding level: binding level between mutant peptide and HLA allele .

         Frequency in the tissue: neoantigen frequency in the samples of specific tumor type.

         Frequency in all samples: neoantigen frequency in all samples.

In the ‘Search‘ page, the gene name is required.

Each cell of the figure stands for a combination of mutation and HLA allele. Only the cell with color (neither gray nor white) represent a potential neoantigen. 

          The information in each colorful cell stands for mutation, HLA allele and the frequency of the neoantigen, respectively. 

          We arrange the result according to the frequencies of HLA alleles and mutations. The frequency of mutation decreased from upper to bottom, and the frequency of HLA allele decreased from right to left. So the most promising potential neoantigens for tumor immunotherapy are listed on the left bottom corner.

In the ‘Validation‘ page

IEDB result: ‘Positvie’ indicates binding, ‘Negative’ indicates non-binding.

IEDB assay ID: The id from IEDB, linked with the detailed information of the result (including reference, experiment assays).



Address:      College of Pharmaceutical Sciences, Zhejiang University, Hangzhou, China.