GPTomics

Research at the intersection of biology and artificial intelligence.

[Jason Chin @infoecho]

Physics ditched the ether and unlocked a deeper understanding of space-time. Genomics still clings to its equivalent: the reference genome. What if ditching it unlocks a deeper understanding of the genome itself?

[dough though @domenjemec]

Part 2 of the series where we look at how physics dealt with absolution position and it's parallel in genomics. On the surface the idea of genomic relativity seems like semantics but we think that it would unlock a new way of looking at the hardest genomics problems.

[GPTomics @gptomics]

We've shown genomic position needs reform. It's based on a broken reference based system. Through special relativity, physics ditched the idea of absolute rest, so @infoecho and @domenjemec propose that we need Genomic Relativity https://t.co/Tn0HY6eaga

[Jason Chin @infoecho]

Having helped bring long-read technology from its dark age to a new dawn by tackling genome assembly, I've been reflecting on what it all means.

I'll never forget the first time I understood the beauty of Einstein's relativity. I hope we can get there with genomics soon.

[dough though @domenjemec]

Physics for a long time focused on swapping out what the reference for object position was, from earth, to the sun, to the ether, and eventually Einstein just removed the need for absolute position. We're seeing a lot of similarity in genomics with position/variants.

[GPTomics @gptomics]

@infoecho and @domenjemec continued to think about how genomics is defined by the reference. They found there was a parallel with physics and its attempts to understand what the center of the universe was https://t.co/lBuXFQBAzo

[Jason Chin @infoecho]

I remembered when I was in Applied Biosystems as a software engineer discussing about "copy number variation". I was so confused about the concept.... 1/N

https://t.co/QFw6ahCukp

[dough though @domenjemec]

Sadly this is the state of a log of genomic sequencing today. Luckily, we have the technology now where we don't need it to be.

[GPTomics @gptomics]

The reference genome has been useful for advancing genomics, but, it's starting to become a systemic limitation. @infoecho and @domenjemec explored what a "variant" actually means and why the relying on references is holding the space back. https://t.co/uOFGP7fQPz

[dough though @domenjemec]

If you've wondered how the layer by layer mechanics of protein folding models work, I created a series of notebooks that walks through each layer. This was way easier to do thanks to @biohub who kept the ESM models open source. https://t.co/6Io9VJnEjo

[GPTomics @gptomics]

@peterottsjo fair enough. There's so much coming into the field the last few months, it's great to see. We launched bioSkills in January fully open source with the hope of seeing it inspire the field to look more at use cases vs tools and the companion evals that focus on clear A|B testing

[GPTomics @gptomics]

@peterottsjo Seems you’re missing BioSkills, a repo of over 500 open source bioinformatics use case focused skills https://t.co/Gr4Nv4eg6K

[dough though @domenjemec]

It was really fun writing this. One of the more interesting things I learned writing this was how Mendel's work remained obscure for over 30 years.

[GPTomics @gptomics]

@domenjemec @infoecho and @BMLauro19 have teamed up to dig into the modern history of comp bio. The first chapter, now available, focuses on how modern statistics and biology co-evolved to form the foundations for comp bio. https://t.co/LZ7fBtWlT2

[Jason Chin @infoecho]

I enjoy using IGV, but I need a more efficient tool for analyzing diploid genomes. So, I developed one for myself. It provides exceptional GPU acceleration, allowing me to uncover all essential information spanning six orders of magnitude. #bioinformatics #genomics #visualization

[dough though @domenjemec]

@AnthropicAI I don't think it's news to anyone in the bio space that disparate data and differing protocols are a huge issue. @GA4GH and others have been trying to solve the FAIR problem for a while.

[dough though @domenjemec]

It was really interesting seeing the insight @infoecho brought about the history of genome assemblers and how they were shaped so much by their different lab dynamics.

[GPTomics @gptomics]

@infoecho @domenjemec reviewed Conway's Law and the history of genome assemblers to see how including agents in academic labs will shape the research software ship. PIs will bifurcate further: either they'll ship more or end up even more removed. https://t.co/gHrjbJT0HB

[GPTomics @gptomics]

While simple, the tool built with this markdown has simplified how research teams keep in sync. Each researcher can pursue their own tangent but stay informed since the LLM Research Manager creates a unified project state contract for all their agents to read and contribute to.

[dough though @domenjemec]

I've been building a tool to enable agents to do research project management: track the evolution of a hypothesis based on changing evidence in a directory/folder. Inspired by @karpathy's LLM-wiki, I decided to deliver it as a gist your agent can implement https://t.co/5dKRz74bLW

[GPTomics @gptomics]

for our next project we're going to try to follow a different pattern of "building software" where we write an idea file. Since our next project will be a tool to help Agents, we think it's natural that your agent build the tool to help it. Stay tuned.