neoFusion: a computational pipeline for gene fusion neoantigen prediction and immunogenic potential assessment

Fusion, which is an important class of somatic mutations, is an ideal source of tumor-derived neoantigens for creating an open reading frame. Given RNAseq sequencing data as input, neoFusion predict and evaluate the immunogenic potential of gene fusion based neoantigen. neoFusion is the first pipeline for predicting and prioritizing fusion neoantigen. Detailed information please refer to citation.

Citation:

The Landscape of Tumor Fusion Neoantigens: A Pan-Cancer Analysis, iScience, 2019.

Dependencies

Required software:

• NetMHCpan 4.0
• OptiType
• netchop
• STAR-Fusion
• [ pepmatch ] module of MuPeXI

Python2 package

 xgboost
 biopython
 scikit-learn
 pandas
 numpy
 subprocess
 multiprocessing
 pickle   

reference

 neoFusion support two mode, denovo and midway. In midway mode, user do not need download reference. In denovo mode,
 please download reference as depicted:  
 The version of STAR-Fusion in our docker is 1.4.0, for compatibility, 
 download GRCh38 reference from (https://data.broadinstitute.org/Trinity/CTAT_RESOURCE_LIB/__genome_libs_StarFv1.3/)
 please refer to STAR-Fusion for more detail.    
 All the other needed materials can be found at data directory   

Installation

Install via Docker, highly recommended

Docker image of neoFusion is available at https://hub.docker.com/r/bm2lab/neofusion/.
if you have docker installed, you can pull the image like so:

docker pull bm2lab/neofusion

Install from source, not recommended

To install neoFusion, you just need download neoFusion.py and all the described dependencies.

Usage

if install from docker:

docker run -it bm2lab/neofusion
python /usr/local/neoFusion.py -h   

if install from source:

please edit the software path in neoFusion.py to the right path in your environment.  

Two mode are provided, denovo and midway. In denovo mode, you should provide fastq files.

python neoFusion.py denovo --left 1.fq  --right 2.fq  --hla  HLA-A02:01  --genome STAR_Fusion/GRCh38/ctat/

In midway mode, just provide fusion proteins in fa format.

python neoFusion.py midway --fusion example_protein.fa --hla HLA-A02:01     

output

The output file "neoscore.txt" contains all putative neoantigens information.

Column	Description
HLA	HLA type
mismatch	mismatch between mtpep and wtpep
MTpep	fusion derived candidiate neopeptide
MTpep_score	predicted score of mtpep output by netMHCpan
MTpep_aff	predicted binding affinity of mtpep output by netMHCpan
MTpep_rank	predicted binding affinity percent rank of mtpep output by netMHCpan
MTpep_comb	Combined score of binding affinity, proteasomal C terminal cleavage, and TAP transport efficiency
WTpep	pepmatch_db_x86_64 extracted normal peptide
WTpep_score	predicted score of wtpep output by netMHCpan
WTpep_aff	predicted binding affinity of wtpep output by netMHCpan
WTpep_rank	predicted binding affinity percent rank of wtpep output by netMHCpan
WTpep_comb	Combined score of binding affinity, proteasomal C terminal cleavage, and TAP transport efficiency
Hydro_Model	peptide immunogenic potential based on amino acid hydrophobicity
R	T cell recognition probability
Score	Immunogenicity score of neoantigens

Contact

Zhiting Wei
[email protected]

Qi Liu
[email protected]

Biological and Medical Big data Mining Lab
Tongji University
Shanghai, China.

neoFusion flowchart