HPIpy

HPIpy is a Python-based standalone package that leverages various sequence-based computational models for the prediction of protein-protein interactions (PPIs) between host and pathogen species.

Table of Contents

• Computational approaches
• Host-pathogen models
• Download Miniconda
• Download the package
• Usage
• Contact Us

Computational approaches

Various sequence-based computational approaches implemented in HPIpy:

• Interolog mapping
• Domain-based
• Phylogenetic profiling
• Gene ontology (GO) semantic similarity

Host-pathogen interaction models

HPIpy supports the following pathosystems for protein-protein interactions prediction.

Model	Host	Pathogen(s)
humanVirus	Human	Human-related virus
humanBacteria	Human	Human-related bacteria
animalPathogen	All animals	Virus, fungi, bacteria
plantPathogen	All plants	Virus, fungi, bacteria

Download and Install Miniconda

If not installed, download and install Miniconda:

• Download the latest version from Miniconda
• To install, execute the command:

bash Miniconda3-latest-Linux-x86_64.sh

Download HPIpy

Download the HPIpy package using one of the below options (1 or 2):

1. Clone this repository:

git clone https://github.com/usubioinfo/hpipy.git

2. Obtain package using link:

wget https://kaabil.net/hpipy/downloads/hpipy.tar.gz
tar -xvzf hpipy.tar.gz

To create the conda environment for HPIpy:

cd hpipy
conda env create -f environment.yml
conda activate hpipy

Usage

To view the "Help" section of HPIpy, run:

python3 -m hpipy --help

Basic usage of HPIpy:

python3 -m hpipy --host exampleData/hostProteins.fasta --pathogen exampleData/pathogenProteins.fasta --computation interolog --model humanVirus

Apart from required arguments (as mentioned above), HPIpy contains several options (see package help), which can be used based on the requirement. Example for advanced analysis:

python3 -m hpipy --host exampleData/hostProteins.fasta --pathogen exampleData/pathogenProteins.fasta \
  --computation interolog \
  --model humanVirus \
  --seq_homology 0.8 \
  --num_threads 20 \
  --network \
  --outputdir samplePPIs \
  --interIdentity 40 60 \
  --interEvalue 1e-15 1e-20 \
  --interCoverage 60

If you want to predict the interactions based on different parameters for prediction, use the --resume_ppis option. The program will not run all the steps but only execute PPIs prediction step.

python3 -m hpipy --host exampleData/hostProteins.fasta --pathogen exampleData/pathogenProteins.fasta \
  --computation interolog \
  --model humanVirus \
  --seq_homology 0.8 \
  --num_threads 20 \
  --network \
  --outputdir samplePPIs \
  --interIdentity 70 80 \
  --interEvalue 1e-25 \
  --resume_ppis

Contact Us

For any queries, contact us at [email protected].