Variants2neoanitgen pipeline
To calculate neoantigen from VCF or MAF (MAF-like) file.
Now pipeline has been implemented and tested is
- MAF or MAF-like file -> MAF to VCF (by maf2vcf.pl) -> annotated VCF (by VEP) -> single sample neoantigen information (by pvacseq) -> summary neoantigen for all samples (by shell script) -> neoantigen quality (by R script, Shell script and Python script which from paper neoantigen fitness model).
Concepts
VCF and MAF files are the most popular text files used in bioinformatics for storing DNA variations.
Tumor antigens (also known as Neoantigens) are those antigens that are presented by MHC class I or MHC class II molecules on the surface of tumor cells. Antigens found only on such cells are called tumor-specific antigens (TSAs) and generally result from a tumor-specific mutation. For human tumors without a viral etiology, novel peptides (neo-epitopes) are created by tumor-specific DNA alterations.
What the pipeline do is provide a quick and easy way to predict neoantigens from variant record files - VCF, MAF or MAF-like files. This pipeline is powered by VEP, pVACseq and vcf2maf toolkit etc..
Note, you should have HLA information file of samples when you want to predict neoantigens.
How it works
pVACseq is a well established tool for predicting tumor-specific mutant peptides (neoantigens). However, it can only accept VCF file as input and generate one directory for one samples. This is not convenient to summary the results and expand this tool. Therefore, I build a pipeline to integrate vcf2maf toolkit with pVACseq, which translate MAF or MAF-like files to VCF firstly, and then call pVACseq to predict neoantigens and finally summary the data.
For now, you can use this pipeline to predict neoantigens by MAF file and sample specific HLA information. VCF to neoantigens and maf-like file to neoantigens are need to be done. Besides, after calling neoantigens, you can compute Neoantigen Quality by NetMHC4.0. The method of neoantigen quality computation is published as A neoantigen fitness model predicts tumour response to checkpoint blockade immunotherapy.
Prerequisites
Two main information you must have before you want to predict neoantigens, one is the patient-specific HLAs and the other is detail information of variants, at least have following columns which can be processed by vcf2maf toolkit.
Chromosome Start_Position Reference_Allele Tumor_Seq_Allele2 Tumor_Sample_Barcode
1 3599659 C T TCGA-A1-A0SF-01
1 6676836 A AGC TCGA-A1-A0SF-01
1 7886690 G A TCGA-A1-A0SI-01
Once you have these two information, you can prepare to configure the pipeline now before you use it.
Install conda
You can download conda by web brower, Anaconda link is https://www.anaconda.com/download/ or curl, wget command tool.
I recommend you install the python2 version conda, it can provide you a python2.7 default environment. Some configuration in our pipeline will use python2.7.
Install conda, run your file use sh
for example
wget -c https://repo.anaconda.com/archive/Anaconda2-5.1.0-Windows-x86_64.exe
sh Anaconda2-5.1.0-Linux-x86_64.shthen follow the commands.
If you find any problem about installing conda, you can google it, the anaconda is very popular, so many problem you encounter basically have been fixed or discussed.
Configuration flows
- Clone Variants2neoanitgen pipeline from Github.
git clone https://github.com/ShixiangWang/Variants2Neoantigen.git
- Install and configure requirements
# install R and biopython
# these used to calculate neoantigen quality
# if your conda default env is python 2.7, just
conda install -c r r-essentials
conda install biopython
# if your conda default env is not python2.7, please create a python2.7 env first, then run commands as above
# conda create --name py2 python=2.7
# create a python 3.5 environment, the pvactools need python 3.5
# muliple bioinformatics tools will be installed in this environment too, like blast, samtools etc.
conda create --name pipeline python=3.5# create an env
source activate pipeline
# install pvactools
pip install pvactools
# install perl and vep
# http://asia.ensembl.org/info/docs/tools/vep/script/vep_download.html#installer
conda install -c conda-forge perl
# configure local-lib, see <https://metacpan.org/pod/local::lib#The-bootstrapping-technique>
wget -c https://cpan.metacpan.org/authors/id/H/HA/HAARG/local-lib-2.000024.tar.gz
tar -zxf local-lib-2.000024.tar.gz
cd local-lib-2.000024/
perl Makefile.PL --bootstrap=~/perl5 # you can change the ~/perl5 to another location for managing perl modules
make test && make install
echo 'eval "$(perl -I$HOME/perl5/lib/perl5 -Mlocal::lib=$HOME/perl5)"' >>~/.bashrc
cpan App::cpanminus
cpanm DBI
cpanm DBD::mysql
# if you encounter error (Ubuntu) like <https://stackoverflow.com/questions/4729722/trying-to-install-perl-mysql-dbd-mysql-config-cant-be-found>, run sudo apt-get install libmysqlclient-dev after sudo apt-get update
cpanm Archive::Extract
cpanm Archive::Zip
# install vep
cd ~
git clone https://github.com/Ensembl/ensembl-vep.git
cd ensembl-vep
perl INSTALL.pl
# select the cache file you want, for example 178 for homo-sapiense GRCh37
export VEP_PATH=$HOME/ensembl-vep
export VEP_DATA=$HOME/.vep
perl INSTALL.pl --AUTO f --SPECIES homo_sapiens --ASSEMBLY GRCh37 --DESTDIR $VEP_PATH --CACHEDIR $VEP_DATA
perl convert_cache.pl --species homo_sapiens --version 92_GRCh37 --dir $VEP_DATA
# add following to you ~/.bashrc
export PERL5LIB=$HOME/ensembl-vep:$PERL5LIB
export PATH=$HOME/ensembl-vep/htslib:$PATH
# cp ExAC_nonTCGA.r0.3.1.sites.vep.vcf.gz* to ~/.vep
# You can download from Variants2Neoantigen/data/ dir or generate it following <https://gist.github.com/ckandoth/d135aec7afd8edd73f572584cbf23f5c> (Just content about this is okay)
# test vep
perl vep --species homo_sapiens --assembly GRCh37 --offline --no_progress --no_stats --sift b --ccds --uniprot --hgvs --symbol --numbers --domains --gene_phenotype --canonical --protein --biotype --uniprot --tsl --pubmed --variant_class --shift_hgvs 1 --check_existing --total_length --allele_number --no_escape --xref_refseq --failed 1 --vcf --minimal --flag_pick_allele --pick_order canonical,tsl,biotype,rank,ccds,length --dir $VEP_DATA --fasta $VEP_DATA/homo_sapiens/92_GRCh37/Homo_sapiens.GRCh37.75.dna.primary_assembly.fa.gz --input_file examples/homo_sapiens_GRCh37.vcf --output_file example_GRCh37.vep.vcf --polyphen b --af --af_1kg --regulatory --custom $VEP_DATA/ExAC_nonTCGA.r0.3.1.sites.vep.vcf.gz,ExAC,vcf,exact,1,AC,AN
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/bioconda/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
# make sure conda-forge is on the top
# then
conda install -c bioconda samtools bcftools ucsc-liftover blast
# vcf2maf tool
cd ~
export VCF2MAF_URL=`curl -sL https://api.github.com/repos/mskcc/vcf2maf/releases | grep -m1 tarball_url | cut -d\" -f4`
curl -L -o mskcc-vcf2maf.tar.gz $VCF2MAF_URL; tar -zxf mskcc-vcf2maf.tar.gz; mv mskcc-vcf2maf-* vcf2maf
# usage:
# perl vcf2maf.pl --man
# perl maf2maf.pl --man
# test vcf2maf
perl vcf2maf.pl --input-vcf tests/test.vcf --output-maf tests/test.vep.maf --ref-fasta ~/.vep/homo_sapiens/92_GRCh37/Homo_sapiens.GRCh37.75.dna.primary_assembly.fa.gz --vep-path=$HOME/ensembl-vep/
perl maf2maf.pl --input-maf tests/test.maf --output-maf tests/test.vep.maf --ref-fasta ~/.vep/homo_sapiens/92_GRCh37/Homo_sapiens.GRCh37.75.dna.primary_assembly.fa --vep-path=$HOME/ensembl-vep
# Installing IEDB binding prediction tools (strongly recommended)
# http://pvactools.readthedocs.io/en/latest/install.html#installing-iedb-binding-prediction-tools-strongly-recommended
# make sure tcsh and gawk are installed
Usage
First add path to your ~/.bashrc and add permission to scripts, so you can run the script from any location.
./add_path.shNew Way
From version 0.2.0, new usage is introduced.
Now you can deploy Variants2neoantigen in two steps:
- Preparation - provide input PATH and specify output PATH, generate configure file (inputArgs.txt)
- Run pipeline - run pipeline, the script will use configuration in file under current directory (generated by preparation part)
The corresponding commands are:
- Preparation
runNEO_prepare [-i <input_file>] [-o <output_dir>] [--HLA <HLA_file>]This process provide input file including MAF and HLA info and specify output directory path.
- Run pipeline
runNEO_runThis process start the pipeline, please remember running runNEO_run in current directory (the path where your configure file inputArgs.txt are).
You also can use this tool as before.
Old Way
Modify argument's template file configureArgs under Variants2Neoantigen directory. Only change the difference between you and this file by following the comments.
# This file is used to configure default parameters of neoantigens computation
# If you do not know what it means, DO NOT change it!!!
# If you want help, free to send email to <[email protected]>
####### <<<<<< Location of Files
##maffiles=$(pwd)/*.maf # location of maf file
##OUTPUT=$(pwd)/neoantigen_results
CACHE_VEP=~/.vep/ # directory path of ensembl vep cache
PATH_FASTA=~/.vep/homo_sapiens/92_GRCh37/Homo_sapiens.GRCh37.75.dna.primary_assembly.fa
# path of reference fasta file
iedb=~/Variants2Neoantigen/data/iedb/iedb.fasta # iedb database used for blastp
##PATH_HLA=~/wangshx/projects/data/ # directory path to patient specific HLA alleles file
####### <<<<<< Location of Softwares, python environment and input arguments
py_env=pipeline # python environment, which can run pvacseq (Python 3.5)
py2_env=py2 # python2.7
maf2vcf=~/vcf2maf/maf2vcf.pl # path of maf2vcf.pl script
PATH_VEP=~/ensembl-vep/ # directory path of ensembl vep software
PATH_VEP_PLUGINS=~/VEP_plugins # directory path of ensembl vep plugins
PATH_MHC=~/MHC # directory path of iedb local installation
vep_run=$PATH_VEP/vep
assembly_version="GRCh37" # genome build version, must be consistent in the analysis
method="NetMHC" # multiple methods separated by space, like "NetMHC PickPocket NetMHCcons NNalign", but multiple method is not recommend
epitope_len="9" # multiple length separated by comma, like "9,10"
Run
runNEO input.maf output_dir input_HLA.tsvHLA input
the input_HLA.tsv should like
Tumor_Sample_Barcode HLA
sample1 HLA-A*30:01,HLA-B*39:01,HLA-B*38:01,HLA-C*12:03,HLA-A*25:01
sample2 HLA-A*31:01,HLA-B*51:01,HLA-A*68:01,HLA-B*35:08,HLA-C*15:02,HLA-C*04:01
sample3 HLA-A*02:01,HLA-B*44:03,HLA-B*07:02,HLA-A*29:02,HLA-C*16:01,HLA-C*07:02
sample4 HLA-B*15:01,HLA-A*02:01,HLA-B*38:01,HLA-C*03:03,HLA-C*12:03,HLA-A*23:01
sample5 HLA-B*44:03,HLA-B*40:01,HLA-C*03:04,HLA-A*24:02,HLA-A*23:01,HLA-C*02:02
sample6 HLA-A*02:01,HLA-B*13:02,HLA-A*30:01,HLA-C*06:02
sample8 HLA-A*68:01,HLA-B*40:01,HLA-C*04:01,HLA-A*68:02,HLA-B*53:01,HLA-C*03:19
two column, the first is the sample id matched with maf file, the second is the HLA, they are tab separated (tsv file, the header can be none).
All neoantigen results are under output_dir/neoantigens/.
License
GPL-3 ©ShixiangWang