This tool allows for the prediction of putative host-pathogen protein-protein interactions based on numerical encoding of physicochemical descriptors
This tool has been tested on Ubuntu (18.04) and MacOS Catalina. Please, make sure to install local blast, R and Zenity.
Install blast locally: bash sudo apt-get install ncbi-blast+
Install R: bash sudo apt install r-base
Install Zenity: bash sudo apt install zenity
To install blast locally follow the instructions in: https://www.ncbi.nlm.nih.gov/books/NBK52640/
Install R package in: https://cran.r-project.org/bin/macosx/
Install Zenity*: brew install zenity
Please, download the repository as a zip file and uncompress it
Download dataset/ and database/ directories* from Zenodo into the main directory where the repository is allocated in your system. Then, unzip the files.
*Dataset and database directories are available under: https://zenodo.org/record/4668840#.YHm-B-0zYVs
In order to run the tool, go to /scripts inside the main directory and run ./main_script.bash
After this, Zenity interface will be prompted and the user will be asked to choose: 1) A host organism (either from the available database in the repository, from uniprot via its taxon ID or from a custom file) 2) A pathogen organism 3) From a set of physicochemical descriptors (either custom or default) to perform the predictive analysis 4) The False discovery rate (this is a measure of astringency, the lower the chosen value, the more astringent the analysis will be) 5) The percentage of physicochemical models that have to agree on a prediction in order to report it in the consensus interactome
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.