GAIAAn agentic bioinformatics solution
GAIA started as a visual way to run real bioinformatics pipelines. Now pivoted to Agentic, combining the power of AI with the precision of traditional bioinformatics.
Where it started
Writing code was never my forte. Now, AI has made it possible to build complex software without writing every line from scratch.
I'm a biologist, not a software engineer. The science was never the blocker; the tooling was. So I decided to build the version I wanted to use.
What I'm building
Modern AI coding tools made it realistic to build something this complex without writing every line of the underlying pipeline logic by hand. The bet was a canvas-based interface. Now introducing the agentic approach.
The system diagram is exposed. Open source from the start — if the gap existed for me, it exists for a lot of other people coming from wet-lab backgrounds, and the fix shouldn't be locked behind one person's repo.
How it works
GAIA is a combination of of Large Language Models, curated and task specifi MCP servere collection containing the indistries most used algorithms and software. We are working to connect these tools into a seamless agentic workflow. Our goal is not to replace humans, but to augment their capabilities. A human-in-the-loop approach ensures that the AI's decisions are aligned with scientific goals and ethical considerations.
- Omics-ready modules — genomics, transcriptomics, proteomics workflows
- AI-assisted configuration — parameter guidance without needing to know defaults
- Open source — fully forkable, extendable by the community
- Community — A collaborative environment for users (researchers, clinicians) to share workflows and insights
Built using AI coding tools throughout — which is also the argument for the project itself: this kind of software is buildable by domain experts, not just engineers, as long as you check every step it hands you.
Current state
Development is paused. I stepped away to build PANOPTES after the Strike liquidation event and haven't returned to GAIA full-time since. The repo is public and the architecture holds up — it's a matter of when, not if, I pick it back up.
The real lesson so far: the hard part isn't the bioinformatics, it's designing an interface abstract, and connecting the analog way wit the agentic way enough to be usable and precise enough to still be scientifically rigorous. That tension is what makes the problem interesting.
View on GitHub