top | item 44695410

(no title)

JayEquilibria | 7 months ago

Maybe I dream of building a competitor and Traumatic Brain Injury is deeply personal to me so I am super biased but it is super "meh" because of Cytoscape.

To me, Cytoscape is simply outdated and an arbitrary constraint on the software, a relic of network science history.

The deeper problem though is how valuable is network data visualization from an epistemological standpoint? Pretty useless from my experiments besides for a kind of scientism performance art. That is not to say scientistic performance art data viz can not be lucrative in a post modern scientism society.

Take this with a grain of salt because biological networks are never going to be my strong point. There is this huge node data viz scaling problem I don't think has been solved and I am kind of waiting for this to be solved with biological network so I can just rip off the idea.

discuss

order

rowanxmas|7 months ago

What about the Cytoscape Desktop metaphor do you not like? As one of the original folks on the project I wrote the early API, and we very specifically made it data-agnostic. So unless you don't want to use nodes and edges (in which case yeah... look elsewhere) I think it's pretty adaptable.

When I was doing grad school work while working on Cytoscape the rendering engine part was only for figure generation. Most of the "real" work was using the networks-subnetworks and doing interesting things with the graph model.

jmg421|7 months ago

Thank you for commenting and for the amazing work you did to get Cytoscape off the ground. If it weren't for your work, nodes.bio would not have been possible!

-John

jmg421|7 months ago

@JayEquilibria- This is a great comment, thanks so much, it helps tremendously. I will reply separately in a more comprehensive way :-).

jmg421|7 months ago

Hey @JayEquilibria — thanks again for your thoughtful comment!

I used to share some of your skepticism — a lot of network visualizations do feel like scientistic performance art. That started to shift for me after discovering NFX and reading The Cold Start Problem by Andrew Chen. I began thinking more seriously about network effects when I applied for my first U.S. patent after leaving JP Morgan Chase, where I spent 14 years in technology, leadership, and innovation. That experience led to the early ideas behind Nodes.bio.

I do rely on Cytoscape.js as a rendering engine — so I want to be transparent about that. But I’ve moved deliberately away from the legacy metaphors and plugin model of desktop Cytoscape. Nodes.bio is browser-native, install-free, and designed for speed, storytelling, and shareability. It's built for researchers, founders, investors, and even patients — not just PhD bioinformaticians.

I understand the “performance art” critique, but I’ve seen what happens when visualization actually delivers insight. One of my early users, Dr. Patrick Sewell, a clinical geneticist, has gone on record saying:

    “Nodes.bio transformed my hand-drawn network into a publication-quality, interactive diagram in minutes—something I simply couldn't have achieved with any other tool.”
That kind of feedback reshaped how I think about the epistemic value of visualization. In a world increasingly shaped by LLMs, I believe pictures that are worth thousands of tokens are becoming essential — not ornamental.

And even from a utilitarian standpoint — forget truth claims, just ask: does this help people make better decisions, faster, with fewer errors? From what I’ve seen: yes. Visual structure helps surface non-obvious connections, prioritize experiments, flag off-target effects, and bridge gaps between data producers and decision makers. That’s reason enough to keep going.

As for scaling: I’m running on AWS ECR/EB, so infrastructure is solid. The real challenge is cognitive scaling — how to keep massive biological networks legible and meaningful. That’s where I’m focusing next.

I’ll also be publishing the Nodes.bio APIs in the near future — because the value here isn’t just in pretty diagrams, it’s in enabling others to build, extend, and integrate with their own data pipelines, dashboards, or discovery workflows.

Appreciate the challenge. If you do decide to build in this space, I’d love to trade notes!