Nextflow Pipeline for Cif/Pdb File Processing

This pipeline automates the process of querying a Neo4j database, downloading structures, rendering videos, and uploading files to an S3 bucket. The pipeline is managed using Nextflow, and requires cypher-shell to interact with the Neo4j database.

Prerequisites

Before running this pipeline, ensure you have the following installed:

1. Neo4j: Follow the instruction to install the docker Neo4j Reactome database, don't forget to run it. If you want to use a Neo4j Deskop modify the address in nextflow.config. to run with docker launch :

   docker run -p 7474:7474 -p 7687:7687 -e NEO4J_dbms_memory_heap_maxsize=8g reactome/graphdb:latest

2. Nextflow: Follow the installation instructions on the Nextflow website. Java is required.
3. Cypher Shell: This is required to run Cypher queries against the Neo4j database. Download it from the Neo4j website.
4. Molstar: Before installing the Mol* packages, be sure to have nodes.js (17+) installed. You must have python and python3 installed. The way to do it depend of your configuration, check on the web. Install all of these packages.

apt-get update && sudo apt-get install -y \
    default-jre \
    wget \
    unzip \
    net-tools \
    ffmpeg \
    pkg-config \
    libx11-dev \
    libxi-dev \
    libcairo2-dev \
    libpango1.0-dev \
    libjpeg-dev \
    libgif-dev \
    librsvg2-dev \
    build-essential \
    libglx-dev \
    libgl-dev \
    libgl1-mesa-glx \
    libgl1-mesa-dri \
    xvfb \
    libsm6 \
    libxext6 \
    libgl1-mesa-dev \
    libosmesa6-dev \
    xorg \
    xserver-xorg \
    libxext-dev \
    libglapi-mesa \
    mesa-utils \

Then you can run :

   cd molstar
   npm install 
   npm run rebuild
   cd ../

5. AWS: Ensure the connection with an AWS s3 server.

brew install awscli
aws credential

7. Before launching the pipeline, you may need to install requests python package.

   python3 -m pip install requests
   (you might need to : --break-system-packages)

or use venv

   python3 -m venv 'path/to/venv'
   source path/to/venv/bin/activate
   python3 -m pip install requests 

Pipeline Parameters

• input: Directory where pre-downloaded AlphaFold .gz files are saved.
• output: Directory where PDB files will be saved (default: "Cif_files").
• cypherScript: Path to the Cypher query script (default: "queryCyph.cyp").
• version: Version of the pipeline (default: "0.1").
• cyphershell_version: Version of the CypherShell (default: "5.21.0")
• address: Address of the Neo4j database (default: "bolt://localhost:7687").

Workflow

The pipeline consists of several processes:

1. Neo4j Query (neo4j)

Queries the Neo4j database using a Cypher script and extracts UniProt IDs.

Output: uniProtID.txt

2. Search and Download Structure (searchAndDownloadStructure)

Searches for PDB structures using the extracted UniProt IDs and downloads them.

Inputs: uniProtID Outputs:

• PDB structure files (${params.output}/${uniProtID}.cif)
• Additional files in ${params.output}/files/

3. Molstar Rendering (molstar)

Renders images of PDB structures using Mol*.

Inputs: PDB structure files (*.cif) Outputs: Rendered images in output_molstar/

4. S3 JSON Upload (s3_json)

Uploads JSON files to an S3 bucket.

Inputs: path_s Outputs: Files uploaded to s3://download.reactome.org/structures/

5. S3 Videos Upload (s3_videos)

Uploads video files to an S3 bucket.

Inputs: path_m Outputs: Files uploaded to s3://download.reactome.org/structures/

Running the Pipeline

1. Configure Nextflow: Ensure Nextflow is properly installed and configured on your system.

2. Prepare Cypher Script: Create or modify the Cypher script (queryCyph.cyp) that contains the query to extract UniProt IDs.

3. Run Nextflow: Execute the following command to start the pipeline:

   nextflow run main.nf

   (in case of error think about the -resume)

Example Cypher Script (queryCyph.cyp)

An example Cypher script to extract UniProt IDs might look like this:

MATCH (n:Protein)
RETURN n.uniProtID

Notes

• Ensure that the Neo4j database is running and accessible at the specified address.
• Make sure AWS CLI is configured with the appropriate permissions to upload files to the specified S3 bucket.
• The molstar process assumes that webm_renderer.js from the Mol* library is available at the specified path.

For any issues or questions, please refer to the Nextflow documentation or the relevant tool documentation.