This is a tool I developed to combine real-time data feedback with bioinformatics workflows. It’s a Node.js server with four long-running threads that provide:
- Sequence visualization via IGV.js
- Automatic indexing of FASTA/BAM/GFF files into FAI/BAI/BGZ and BGZ into CSI/TBI
- Parallel BLASTn searches with direct linking of hits to IGV applets
- URL-based uploads for remote data sources
Integrated LLM agents (GPT-4o, Kong API, etc.) that can dynamically modify the interface, re-sort BLAST tables, summarize sequences, or analyze what’s displayed. Multiple agents can run in one window.
To start, run:
/bin/bash start.sh
The repository includes a full tutorial video that walks through setup from scratch inside a clean Ubuntu 24 container. Xpress-Bio handles multi-channel WebSockets, Server-Sent Events, and real-time client updates so you can focus on your own pipelines. You can spawn additional worker threads to run custom jobs and stream progress back to the user interface.
Samtools, htslib, bcftools, and blastn are already bundled, and the environment is extendable.
The project’s still evolving, feedback helps shape where it goes next. If there’s functionality you’d like to see, let me know, I can try to add it.
Here’s a quick screenshot for context:
