Brian Naughton
@btnaughton.bsky.social
genetics/data/programming. ex-Hexagon, ex-Stanford ex-23andMe ex-TCD http://blog.booleanbiotech.com 🇮🇪
I decided to experiment with my own (small) protein design competition!
It's specifically to test how well VHH pipelines work for binder design, without the usual careful tuning.
You submit a modal script, I run it and test on benchbb.bio
blog.booleanbiotech.com/vhh-competit...
It's specifically to test how well VHH pipelines work for binder design, without the usual careful tuning.
You submit a modal script, I run it and test on benchbb.bio
blog.booleanbiotech.com/vhh-competit...
October 16, 2025 at 4:54 PM
I decided to experiment with my own (small) protein design competition!
It's specifically to test how well VHH pipelines work for binder design, without the usual careful tuning.
You submit a modal script, I run it and test on benchbb.bio
blog.booleanbiotech.com/vhh-competit...
It's specifically to test how well VHH pipelines work for binder design, without the usual careful tuning.
You submit a modal script, I run it and test on benchbb.bio
blog.booleanbiotech.com/vhh-competit...
Another promising VHH model just dropped today from Manifold Bio. This one builds on BindCraft and ColabDesign. MIT license!
www.biorxiv.org/content/10.1...
The success rates appear to be lower than other tools, but this is highly target-dependent.
www.biorxiv.org/content/10.1...
The success rates appear to be lower than other tools, but this is highly target-dependent.
September 29, 2025 at 7:03 PM
Another promising VHH model just dropped today from Manifold Bio. This one builds on BindCraft and ColabDesign. MIT license!
www.biorxiv.org/content/10.1...
The success rates appear to be lower than other tools, but this is highly target-dependent.
www.biorxiv.org/content/10.1...
The success rates appear to be lower than other tools, but this is highly target-dependent.
A REALLY nice use of nanobanana (Gemini) is cleaning up blurry old images. Of course it depends what you ask for, but it is amazing at keeping the text and content the same.
September 19, 2025 at 1:53 PM
A REALLY nice use of nanobanana (Gemini) is cleaning up blurry old images. Of course it depends what you ask for, but it is amazing at keeping the text and content the same.
We decided to try to 3d print custom chess pieces.
The workflow of asking gemini to iterate on a design image, then uploading that image to adam.new worked amazingly well. 5 minutes work!
Left is the gemini image and right is the mesh.
The workflow of asking gemini to iterate on a design image, then uploading that image to adam.new worked amazingly well. 5 minutes work!
Left is the gemini image and right is the mesh.
September 13, 2025 at 10:14 PM
We decided to try to 3d print custom chess pieces.
The workflow of asking gemini to iterate on a design image, then uploading that image to adam.new worked amazingly well. 5 minutes work!
Left is the gemini image and right is the mesh.
The workflow of asking gemini to iterate on a design image, then uploading that image to adam.new worked amazingly well. 5 minutes work!
Left is the gemini image and right is the mesh.
I was reminded about Blackett's War by a mention on the @patio11 podcast. Terrific book about operations research / probability theory applied to real world problems.
Lots of anecdotes like this...
Lots of anecdotes like this...
September 6, 2025 at 3:05 PM
I was reminded about Blackett's War by a mention on the @patio11 podcast. Terrific book about operations research / probability theory applied to real world problems.
Lots of anecdotes like this...
Lots of anecdotes like this...
Ibex from Prescient Design arxiv.org/abs/2507.09054 github.com/prescient-de...
- antibody folding with performance like Boltz/Chai but up to 100X faster
- does not fold the complex, just the antibody
- predicts holo and apo forms
- sadly, model weights are not freely available
- antibody folding with performance like Boltz/Chai but up to 100X faster
- does not fold the complex, just the antibody
- predicts holo and apo forms
- sadly, model weights are not freely available
August 13, 2025 at 4:31 PM
Ibex from Prescient Design arxiv.org/abs/2507.09054 github.com/prescient-de...
- antibody folding with performance like Boltz/Chai but up to 100X faster
- does not fold the complex, just the antibody
- predicts holo and apo forms
- sadly, model weights are not freely available
- antibody folding with performance like Boltz/Chai but up to 100X faster
- does not fold the complex, just the antibody
- predicts holo and apo forms
- sadly, model weights are not freely available
Asking Claude to research binder designs metrics and it gave me my own blogpost back!
Time to start writing for the AIs? marginalrevolution.com/marginalrevo...
Time to start writing for the AIs? marginalrevolution.com/marginalrevo...
July 19, 2025 at 2:13 PM
Asking Claude to research binder designs metrics and it gave me my own blogpost back!
Time to start writing for the AIs? marginalrevolution.com/marginalrevo...
Time to start writing for the AIs? marginalrevolution.com/marginalrevo...
Vibe coded a run progress simulator with Gemini Canvas.
It worked great, until the context rot doom loop set in, and now it's difficult to fix the remaining minor bugs.
Still very impressive!
hgbrian.github.io/run_progress...
It worked great, until the context rot doom loop set in, and now it's difficult to fix the remaining minor bugs.
Still very impressive!
hgbrian.github.io/run_progress...
June 23, 2025 at 2:56 PM
Vibe coded a run progress simulator with Gemini Canvas.
It worked great, until the context rot doom loop set in, and now it's difficult to fix the remaining minor bugs.
Still very impressive!
hgbrian.github.io/run_progress...
It worked great, until the context rot doom loop set in, and now it's difficult to fix the remaining minor bugs.
Still very impressive!
hgbrian.github.io/run_progress...
BindEnergyCraft claims to improve BindCraft performance using an energy-based objective.
"Code will be released soon."
"Code will be released soon."
May 30, 2025 at 7:43 PM
BindEnergyCraft claims to improve BindCraft performance using an energy-based objective.
"Code will be released soon."
"Code will be released soon."
This is what Claude gave me. It seems to have done a pretty good job, but 200k is maybe too many clinical trials! There are a ton of "clinical studies" being done but I don't know what that means.
April 25, 2025 at 1:43 PM
This is what Claude gave me. It seems to have done a pretty good job, but 200k is maybe too many clinical trials! There are a ton of "clinical studies" being done but I don't know what that means.
There are now a trio of great pieces covering the rapid rise of Chinese biotech:
- atelfo.github.io/2024/12/20/w...
- centuryofbio.com/p/commoditiz...
- www.asimov.press/p/china-trials
- atelfo.github.io/2024/12/20/w...
- centuryofbio.com/p/commoditiz...
- www.asimov.press/p/china-trials
April 24, 2025 at 1:35 PM
There are now a trio of great pieces covering the rapid rise of Chinese biotech:
- atelfo.github.io/2024/12/20/w...
- centuryofbio.com/p/commoditiz...
- www.asimov.press/p/china-trials
- atelfo.github.io/2024/12/20/w...
- centuryofbio.com/p/commoditiz...
- www.asimov.press/p/china-trials
sciencemachine.ai looks really interesting (default example analysis below)
I am not sure how uploading to the browser can work for real sequencing data though. Maybe they will take a bucket url or something.
I am not sure how uploading to the browser can work for real sequencing data though. Maybe they will take a bucket url or something.
April 23, 2025 at 2:07 PM
sciencemachine.ai looks really interesting (default example analysis below)
I am not sure how uploading to the browser can work for real sequencing data though. Maybe they will take a bucket url or something.
I am not sure how uploading to the browser can work for real sequencing data though. Maybe they will take a bucket url or something.
I did not know @plasmidsaurus.bsky.social is doing eukaryote nanopore sequencing now. The price is good, especially $50 extra for extraction.
Unfortunately you cannot yet send any tissue, only certain blood and cell culture.
Unfortunately you cannot yet send any tissue, only certain blood and cell culture.
April 22, 2025 at 5:12 PM
I did not know @plasmidsaurus.bsky.social is doing eukaryote nanopore sequencing now. The price is good, especially $50 extra for extraction.
Unfortunately you cannot yet send any tissue, only certain blood and cell culture.
Unfortunately you cannot yet send any tissue, only certain blood and cell culture.
Since BoltzDesign1 just got published, people might also be interested in github.com/escalante-bi... by @nboyd.bsky.social
It has a nicely flexible system for including custom losses. It's also surprisingly little code on top of Boltz (and the Jaxified Joltz).
It has a nicely flexible system for including custom losses. It's also surprisingly little code on top of Boltz (and the Jaxified Joltz).
April 10, 2025 at 12:00 AM
Since BoltzDesign1 just got published, people might also be interested in github.com/escalante-bi... by @nboyd.bsky.social
It has a nicely flexible system for including custom losses. It's also surprisingly little code on top of Boltz (and the Jaxified Joltz).
It has a nicely flexible system for including custom losses. It's also surprisingly little code on top of Boltz (and the Jaxified Joltz).
Checking into options for home lab testing. A lot of places are now offering the Tasso device or similar for home collection:
- mylabsdirect.com/product/comp...
- rythmhealth.com
- siphoxhealth.com
- www.functionhealth.com
- www.walkinlab.com
- mylabsdirect.com/product/comp...
- rythmhealth.com
- siphoxhealth.com
- www.functionhealth.com
- www.walkinlab.com
March 31, 2025 at 4:52 PM
Checking into options for home lab testing. A lot of places are now offering the Tasso device or similar for home collection:
- mylabsdirect.com/product/comp...
- rythmhealth.com
- siphoxhealth.com
- www.functionhealth.com
- www.walkinlab.com
- mylabsdirect.com/product/comp...
- rythmhealth.com
- siphoxhealth.com
- www.functionhealth.com
- www.walkinlab.com
Kurzweil still takes Resveratrol??
February 20, 2025 at 4:27 AM
Kurzweil still takes Resveratrol??
Interesting case study:
Spontaneous tumor regression following COVID-19 vaccination (2021)
jitc.bmj.com/content/10/3...
Spontaneous tumor regression following COVID-19 vaccination (2021)
jitc.bmj.com/content/10/3...
February 11, 2025 at 5:09 PM
Interesting case study:
Spontaneous tumor regression following COVID-19 vaccination (2021)
jitc.bmj.com/content/10/3...
Spontaneous tumor regression following COVID-19 vaccination (2021)
jitc.bmj.com/content/10/3...
February 9, 2025 at 9:15 PM
Excerpt from "Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial", BMJ, 2018 pubmed.ncbi.nlm.nih.gov/30545967/
February 2, 2025 at 4:44 AM
Excerpt from "Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial", BMJ, 2018 pubmed.ncbi.nlm.nih.gov/30545967/
searching for pdb entries and google randomly did this
January 23, 2025 at 7:25 PM
searching for pdb entries and google randomly did this
January 22, 2025 at 5:50 PM
This excerpt from the new
@owl_posting
article reminded me of an interesting wikipedia page.
"2. Extended half-life: Antibodies stay in your bloodstream for weeks vs < 1 day for small molecules. This means less frequent dosing."
@owl_posting
article reminded me of an interesting wikipedia page.
"2. Extended half-life: Antibodies stay in your bloodstream for weeks vs < 1 day for small molecules. This means less frequent dosing."
January 9, 2025 at 5:29 PM
This excerpt from the new
@owl_posting
article reminded me of an interesting wikipedia page.
"2. Extended half-life: Antibodies stay in your bloodstream for weeks vs < 1 day for small molecules. This means less frequent dosing."
@owl_posting
article reminded me of an interesting wikipedia page.
"2. Extended half-life: Antibodies stay in your bloodstream for weeks vs < 1 day for small molecules. This means less frequent dosing."
The claims in that abstract were even stronger, so I do wonder if they pulled back a bit.
January 4, 2025 at 4:10 PM
The claims in that abstract were even stronger, so I do wonder if they pulled back a bit.
Impressive claims from the recent EvolveX paper www.biorxiv.org/content/10.1...
unfortunately it relies on a commercial tool, FoldX, so it's hard to test out
unfortunately it relies on a commercial tool, FoldX, so it's hard to test out
January 3, 2025 at 4:38 PM
Impressive claims from the recent EvolveX paper www.biorxiv.org/content/10.1...
unfortunately it relies on a commercial tool, FoldX, so it's hard to test out
unfortunately it relies on a commercial tool, FoldX, so it's hard to test out