NET-evaluation

A package for network evaluation, we provide three independent evaluation metrics: (1) binding, (2a) Gene Ontology (GO), (2b) GO-directness, and (3) PPI. For a detailed description of the metrics, please refer to our NetProphet3 paper [to provide the link once published]. The easiest way to run this package is with the singularity container provided below. The advantage of the container is that there is no need for installing any dependency software except for the singularity software itself. Otherwise, users can still run this package the old way by installing all dependencies (which can be tedious).

I. The easiet, with the Singularity container

1. Install Singularity and download the container

• Refer to Singularity website to install singularity >= 3.5.2
• Pull the singularity container: singularity pull library://dabid/default/s_neteval:sha256.5bbd71f6303bc98d302e089e50bdcc91a3f92e0adf723883ed5a24d41a5a469f
• Rename the container with a shorted name, so it is easier to call: mv s_neteval_sha256.5bbd71f6303bc98d302e089e50bdcc91a3f92e0adf723883ed5a24d41a5a469f.sif s_neteval.sif

2. Clone NET-evaluation repository

git clone https://github.com/BrentLab/NET-evaluation.git

3. Run Commands

There are four different evaluation commands for (1) binding, (2) GO, (3) GO-directness and (4) PPI. There is a sample code for running each of the commands in toy_example/run.sh. Below, we only explain the binding command in details, but the other commands follow the same principles.

• For help:

p_src_code="/path/of/code/that/was/cloned/from/github/repository/"

${p_src_code}binding -h # for binding
${p_src_code}go -h  # for GO
${p_src_code}go_directness -h  # for GO-directness
${p_src_code}ppi -h  # for ppi

• For running the binding command:

${p_src_code}binding \
    --p_in_net "${p_src_code}toy_example/zev.tsv" \
    --nbr_top_edges 50 \
    --nbr_edges_per_threshold 5 \
    --p_binding_event "${p_src_code}metadata/yeast/reg_target_cc_exo_chip_exclusive.txt" \
    --p_out_eval "/path/of/output/file.tsv" \
    --flag_singularity "ON" \
    --p_singularity_img "/path/of/singularity/container/s_neteval" \
    --p_singularity_bindpath "/home/"

II. The more advanced, by installing all dependencies

1. Install required packages

• Install anaconda3 and create environment and call it neteval. Refer to anaconda website on how to install anaconda
• Install python3.6 and pandas within the environment neteval
• For GO analysis, install GO-Term-Finder v0.86
  • GO analysis, we need gene ontology files, specifically p_gene_association (gaf) and p_gene_ontology (obo) files can be downloaded from gene ontology website.
• For running the binding command outside the singularity container:

${p_src_code}binding \
    --p_in_net "${p_src_code}toy_example/zev.tsv" \
    --nbr_top_edges 50 \
    --nbr_edges_per_threshold 5 \
    --p_binding_event "${p_src_code}metadata/yeast/reg_target_cc_exo_chip_exclusive.txt" \
    --conda_env "np3" \
    --flag_debug "OFF" \
    --p_out_eval "/path/of/output/file.tsv" \
    --flag_singularity "OFF" \