Filter neoepitopes based on threshold from Wells et al. (2020)

Clone the repository:

git clone https://github.com/tanyaphung/neoantigens_prioritization.git

Prequisites

• Python version 3.6.11
• Pandas version 1.1.3

Directory structure

• Each sample has its own directory, for example: patient_3466 (under the directory example_data)
  • Sub-directories inside the sample directory are the directories of the HLAs, for example: HLA-A01-01. The directory name of the HLA must match with the hla file for use as argument to the script.
    • Sub-directories inside the HLA directory are called 9_mers or 10_mers or 11_mers
    • Inside sub-directory 9_mers (or 10_mers), there have to be 2 files:
      • netmhc.xsl: output from netMHCpan4.0. Modify line 90 if your file is not in this format.
      • netmhcstab.xsl:output from netMHCpanstab. Modify line 91 if your file is not in this format.

Usage:

python neoepitope_prediction.py --hla_types_fn example_data/patient_3466/hla.txt --quant_fn example_data/patient_3466/3466_quant.sf --sample_id patient_3466 --mers 9 --data_dir example_data/

• This script outputs a file called filtered_neoepitopes.tsv inside the sample directory

If there's no TPM file available:

• If you don't have RNAseq data and therefore do not have the TPM file, you don't have to provide any input to quant_fn. Instead, this script will filter based on MHC binding affinity and MHC stability only.

python neoepitope_prediction.py --hla_types_fn example_data/patient_3466/hla.txt --sample_id patient_3466 --mers 9 --data_dir example_data/

• This script outputs a file called filtered_neoepitopes_no_tpm.tsv inside the sample directory

Required inputs

• --hla_types: this is a file listing each HLA type per line. The name of the HLA type must match with the subdirectory inside the sample directory
• --sample_id: sample id must match the sample directory
• --mers: Enter a number (eg: 9) or a list of numbers 9,10,11
• --data_dir: Enter the parent directory that hosts all the data. For example, because the directory patient_3466 is under the directory example_data, give the path to the example_data directory.

Optional inputs

• --quant_fn: give the path to the output from salmon
• binding_threshold: Default is 34 nM (Wells et al. 2020). Input a different number if you want to change the default threshold
• stability_threshold: Default is 1.4h (Wells et al. 2020). Input a different number if you want to change the default threshold
• tpm_threshold: Default is 33TPM (Wells et al. 2020). Input a different number if you want to change the default threshold