toniher/seqservice repository overview

⁠seqservice

NodeJS express API for handling biological sequences analyses

⁠INSTALLATION

⁠Server configuration

• Dependencies

  • NCBI-BLAST
  • Samtools (optional)
  • HMMER (optional)

• Assuming we define /seqservice as basepath, and we keep 10030 port in config.json

⁠Apache

# Reverse proxy seqservice
RewriteEngine On
RewriteRule ^/seqservice/(.*)$ http://localhost:10030/seqservice/$1 [P]

ProxyPass        /seqservice/socket.io http://localhost:10030/seqservice/socket.io
ProxyPassReverse /seqservice/socket.io http://localhost:10030/seqservice/socket.io

⁠NGINX

Check example nginx.conf file

Info about WebSocket and NGINX: https://www.nginx.com/blog/websocket-nginx/⁠

⁠Application start

From the directory of the application

npm install
npm run build
NODE_ENV=production node index.js config.json

FUN!

⁠Docker installation

From this repository:

docker build -t seqservice .
docker run -p 10030:10030 -v /path/to/db:/data/db/seqservice -v /path/to/config.json:/data/soft/config.json -d seqservice

Open your browser at http://localhost:10030/seqservice/

In any case, adapt database and config.json path and contents to fit your needs

⁠TODO

• Webpack: Consider putting manifest for hashed version
• Webpack and npm run: enable proper PRODUCTION and DEVELOPMENT behaviour
• Allow single-hit accompanying analyses (e.g. ProtLoc or GO)
• Generate alignment from results at user's choice
• Start PSIBLAST from user choice of sequences
• Include more HMMER options
• Download results in different formats (CSV, etc.)
• Allow syncing/saving with a DB (e. g. CouchDB)
• Plug a task queue. By using https://github.com/paulmillr/chokidar⁠ and maybe Nextflow
• Download input databases if needed (through CLI or web interface)
• Plug an embeddable Genome Browser (e. g. igv.js)
• More coherent system of keeping params in configuration for more different analyses
• Avoid calling name API if no names