Surge Biswas
@SurgeBiswas
Followers
3K
Following
3K
Statuses
842
Scientist, co-founder Nabla Bio, dog dad
Joined September 2013
Incredibly excited to share new results from Nabla Bio where we show we can design antibodies de novo for use in therapeutic discovery. We introduce JAM, an AI system we’ve developed to design de novo antibodies with good affinities, early stage developability, and function. We’re seeing success across a range of soluble and hard-to-drug membrane proteins including a Claudin and a GPCR. We’ve extensively tested these designs in our wet lab and included detailed data and controls, providing the first clear demonstration of how de novo design could expand the scope and efficiency of therapeutic antibody discovery. 🧵 Technical report: Blog:
11
63
240
RT @CJHandmer: We're unaffected at @TerraformIndies by grant freezes because despite spending years asking DoE to help out our process of i…
0
49
0
Thread summarizing the results:
Incredibly excited to share new results from Nabla Bio where we show we can design antibodies de novo for use in therapeutic discovery. We introduce JAM, an AI system we’ve developed to design de novo antibodies with good affinities, early stage developability, and function. We’re seeing success across a range of soluble and hard-to-drug membrane proteins including a Claudin and a GPCR. We’ve extensively tested these designs in our wet lab and included detailed data and controls, providing the first clear demonstration of how de novo design could expand the scope and efficiency of therapeutic antibody discovery. 🧵 Technical report: Blog:
0
0
2
@owl_posting @fran_anastas and I have had good luck getting science art printed by amazon photos. Might give it a try, especially if it helps simplify shipping to customers
1
0
5
If you're at BioLogic in San Diego next week, check out Adithya's talk on our work in de novo antibody design!
0
5
8
great and thorough post on one way we @nablabio are designing antibodies to multipass membrane proteins, a high-impact but difficult to drug class of proteins
Better antibodies by engineering targets, not engineering antibodies (Nabla Bio) another startup coverage article! the first one ive written over an antibody startup the interesting part about them? how they think about targets, not antibodies
0
3
17
RT @PearlF: AI investor @dreidco of @ZettaVentures(@NOETIK_ai & @nablabio in portfolio) posted ahead of #JPM25 on what he's paying attentio…
0
3
0
RT @PearlF: Is there a renaissance in GPCR drug development—is the question explored in Nature piece. Novartis president of biomedical rese…
0
1
0
RT @owl_posting: A primer on ML in cryo-electron microscopy (cryo-EM) confused about cryo-EM??? i explain why pe…
0
25
0
More than time/cost, a better question is what problems does AI antibody design enable that's hard/near impossible to solve with conventional approaches. That's where companies should spend their time Atomic precision is a key limiter for many tx applications (e.g. most epitopes not functionally relevant, binding limited extracellular epitopes on a GPCR, high on-target selectivity, etc), and something AI uniquely offers. We've made progress here at @nablabio. Integration of AI and wet-lab has been critical to get it to work in practice.
1
0
3
@HongjiangM @anthonygitter Ah, in this case we do not have a native structure, so JAM is taking only the sequence of Target X as input.
1
0
1
I'm a little confused by how we're using "a lot" here. Are you saying this is a lot of designs to experimentally test, in which case my answer is "not via yeast display". In case it wasn't clear in the paper we pool all 6K designs into a single epPCR reaction, and get ~980M transformants. So construction of the library is not too hard either. Or that it's "a lot of sequences screened, so the affinity results aren't so surprising." To that I'd say people construct much larger naive antibody libraries (10^11-10^12) to achieve a similar number of binders and affinities we're seeing in here. So it's roughly a 10-100x improvement in that regard.
1
0
0
RT @radicalvcfund: In this week's #RadicalReads, @SurgeBiswas, CEO and Co-Founder of @nablabio, shares a breakthrough in de novo antibody d…
0
1
0
@zenbrainest We haven't so far, but definitely worth a conversation. Sending a note to your med unc email, but lmk if there's a better way to get in touch. We are grateful for your lab's work and resources
0
0
0
good questions from @anthonygitter and our answers in the thread below
@SurgeBiswas Congrats @SurgeBiswas, this looks exciting after a quick first pass. I have a few initial questions: - Target X is absent from the PDB. How structurally similar is it to the closest structures that are in the PDB per Foldseek or another tool?
0
0
2