JURA Bio, Inc. (@jura_bio) header
JURA Bio, Inc. (@jura_bio) profile picture

We sample from our generative biological sequence models at peta-scale. Read about our work at https://nitter.cf/t.co/j5R4jau8qI

BOS | BSL | OSL | KBH | LGA
Joined August 2020
Photo by JURA Bio, Inc. (@jura_bio)Photo by JURA Bio, Inc. (@jura_bio)Photo by JURA Bio, Inc. (@jura_bio)Photo by JURA Bio, Inc. (@jura_bio)Photo by JURA Bio, Inc. (@jura_bio)
Super nice! Thanks for the careful reading, Jonas, and for working out some diagrams! I like your illustrations.
Replying to @qbilius
Now that in silico design has become much more effective, the bottleneck has shifted to physically building and testing the variants. The two papers I chose for the inaugural discussion, Variational Synthesis and LIFT (@EliWeinstein6 and @jura_bio), tackle exactly that. 3/5
1
9
1,024
JURA Bio, Inc. retweeted
Now that in silico design has become much more effective, the bottleneck has shifted to physically building and testing the variants. The two papers I chose for the inaugural discussion, Variational Synthesis and LIFT (@EliWeinstein6 and @jura_bio), tackle exactly that. 3/5
1
1
1
4
1,306
We are hiring at @jura_bio both on the ML team (world) and the wetlab team (Boston). Please get in touch if you've been thinking about it. ⏩[email protected]
4
15
1
120
16,961
Curious about @jura_bio work combating multidrug-resistant threats with Walter Reed’s extraordinary physicians and researchers? I’ll share more in a seminar this Thursday at the University of Minnesota’s Dept. of Immunology. Online and open to all; DM if you'd like a link.
2
21
835
Missing @jura_bio — scaling frontier AI in vitro. The world’s largest collection of protein protein interactions. Foundation models trained entirely in the lab.
1
1
5
300
The physical world is exponentially more scalable than the in silico methods that try to recapitulate it. Come learn about @jura_bio: we compute in chemistry because the natural substrate of biological machine intelligence has never been silicon.
Verification is expensive because it's bounded by the physical world ie. what you can glean from a biological assay. Whoever figures out how to scale beyond the physical world without excessively compromising on fidelity and resolution will change the world.
1
4
33
1,554
RT @lizbwood: This paper (@KasperKrunderup and @EliWeinstein6) is a tour de force is experimental codesign, and shows exponential speedups…
This quoted post is unavailable.
1
11
JURA Bio, Inc. retweeted
It's been an exciting week at @jura_bio! We got to share a significant update to Variational Synthesis. cVS gives us more control and flexibility in designing and synthesizing large-scale DNA libraries. As always, @lizbwood wrote a great intro!
Introducing: continuous variational synthesis 📄 jurabio.com/s/cvs.pdf We’re pleased to announce a new advance in our ability to synthesize generative model-designed DNA sequences at petascale.
1
2
3
428
JURA Bio, Inc. retweeted
First you could train an in vitro generative model – now, thanks to our guest star player @AlanNawzadAmin, you can fine-tune it as well!
Introducing: continuous variational synthesis 📄 jurabio.com/s/cvs.pdf We’re pleased to announce a new advance in our ability to synthesize generative model-designed DNA sequences at petascale.
2
3
423
A common misunderstanding. Instead, the model can capture full conditionals -- and then sample in DNA. If the Ab CDR3 example is too confusing (because of the apparent flexibility of that region), then look to the enzyme & human peptide examples from both papers.
I'm in the same boat as Sam. My naive understanding is that it's essentially a fancy NNK library. That makes sense for applications like antibody CDR generation, where the positions can be treated as independent. But in most sequences the probabilities are conditional.
1
3
25
4,848
To be fair to you -- this is a criticism we get a lot. But I have never understood it. To me it's like saying: You developed a 2,000,000,000-megawatt reactor at 1-trillionth the old cost of energy! that's 100x the total world's energy needs! Why'd you do that?
2
10
1,202
That's right. It's just like -- let's design 10^16 sequences and have them all be made. What has always been surprising to me is when these realized designs outperform in-silico-only models, as in the antibody CDR3 (vs. IgLM) task in rdcu.be/e8zUb.
Replying to @anshulkundaje
Yeah I just really don’t understand the use case even. Like I am used to thinking of it as I design some sequences and then they get made. And sure you can have a model design the sequences. But this seems like it is something else that I have not grokked?
3
5
57
12,059
Very pleased to share that arxiv.org/abs/2512.15984v1 -- a method to pack and extract more information from high throughput experiments, using variational synthesis -- was accepted at NeurIPS. You can also read the blogpost: jurabio.com/blog/lift
2
8
44
1,817
JURA Bio, Inc. retweeted
Also @AlanNawzadAmin is one of the most creative computational scientists in bio. Love his work.
2
4
11
1,940
JURA Bio, Inc. retweeted
This looks phenomenal! I think Jura Bio's tech is under appreciated.
Introducing: continuous variational synthesis 📄 jurabio.com/s/cvs.pdf We’re pleased to announce a new advance in our ability to synthesize generative model-designed DNA sequences at petascale.
6
9
73
17,412
JURA Bio, Inc. retweeted
A few years ago, @EliWeinstein6 and @jura_bio wrote the output of a DNA synthesizer as a probability distribution – a family of generative models that you can sample from in real life, Variational Synthesis (VS). But fitting VS models is brutal. Now we’ve fixed that! 1/6
1
14
74
5,087
Today we're introducing continuous variational synthesis. Spearheaded by @AlanNawzadAmin, cVS brings to manufacturing-aware architectures more of what we expect from modern generative model architectures, including pretraining and fine-tuning, and flexibility wrt hardware.
Introducing: continuous variational synthesis 📄 jurabio.com/s/cvs.pdf We’re pleased to announce a new advance in our ability to synthesize generative model-designed DNA sequences at petascale.
1
4
524
I am once again urging everyone to use variational synthesis for problems like this. (Thanks for citing us, Bruce!)
The cost of gene synthesis prevents machine-learning guided directed evolution from reaching it's full potential. A very nice perspective by the inimitable Bruce Wittmann.
1
5
60
6,717
I am grateful that Chinese coverage of an American company is worth reporting and sharing, @SynBioBeta. But why not cover us yourselves sometime?
1
5
237