MolBioMike

Exited to try out my new plasmids from @ATinyGreenCell Thank you Binomica Labs!

You can see the patent cliff here in action Revenues in the white bars for most major companies imply a 50-80%+ drop in the next few years Someone is going to have to fund the next generation of biotech startups for drug discovery so these mature companies can consume and commercialize them Should be an interesting few years imo This is why some of these names are trying to diversify their portfolios into Chinese drugs (particularly bsmAbs or other modalities) There aren’t enough competitive drugs coming out of the west. Anyone can make a pretty good bispec or ADC now it seems. China can make hundreds of candidates allowing big Pharma to have their pick of the litter. The west should focus on what we have always been good at - innovation and new platforms We may need to let the traditional paths to drugs get consumed by less innovative rival nations. The innovation gap in biotech is closing quickly and could look like the AI gap (K2 vs GPT) if we don’t do something about it right now.

@biocheMichael Yo-yo trip to SF

Back in Boulder county 🛬 Clients lab is live including cell culture 🔬 Science exchange is finally approved 🔑 Top three Pharma client unblocked 🔓 Setting up the storefront 🏭 Work on the IP filing 🪺 Back2SF Monday 🛫

@rlacombe Yet another demonstration that self is on a continuum!

You thought biology was weird? It's even weirder. Your cells are out there literally playing tag.

@biocheMichael Claude be making jokes.

Last piece of the stack is allegedly training.

@SynBio1 New bio company idea: genetically engineer humans to become angels 👼

About 7% of deaths are caused by accidents, not disease, so if we are serious about genetically engineering humans to extend lifespan we'll eventually need to redesign the body to better resist drowning poisoning and impact trauma

@ChrisHayduk How long does it take for you to run one experiment? And how are you handling memory and context over the research horizon?

GPT 5.5 is an effective autoresearcher in structural biology! I've had goal mode running for over 150 hours straight, looking for topologically inspired architectural changes to improve the performance of AlphaFold2. Performance is strong and improving!

@ATinyGreenCell I’ve literally just held the electrodes against the cuvette with my gloved hand to get around this problem. Biorad device not this particular one.

Wild to crave product differentiation so badly you make it incompatible with popular gear. The cuvette body is standard but the cap doesn't fit this classic Eppendorf electroporator... Why is misanthropy such a popular tech trend?

@design_proteins @OmicsOmicsBlog The Bagel paper does this but there’s no lab validation on it yet. Was thinking of trying to clone his seqs and give them a try

@MolBioMike @OmicsOmicsBlog Sounds awesome, would be really cool to make mini polymerases

fascinating technology, specific de novo DNA binders enable: - PAM independent targeted genome modification - synthetic transcription factors - sequence blockade therapeutics - synthetic regulatory circuits - DNA biosensors and custom diagnostics biorxiv.org/content/10.648…

Dude I’m so >fucking< psyched about these new models I have built and am testing in silico. After >>100,000 designed proteins using OS stacks for clients and myself I have gone my own way and built new combinations of architectures to produce something totally different. Something like 13 different pivots (🫠) - have the logs to prove it >50 cloud training runs to get here (I learned so much along the way) I built: Something like what I wished existed already Something like a hybridoma on a chip Something Pharma grade imo First class features I hand built into the stack and curated from my real drug discovery experiences: 1. Antigen awareness across the model stack - multiple structure and sequence models wired together intelligently 2. Real developability awareness - no crazy WFY fractions or poly A mess 3. Real frameworks - over 10k real frameworks from real biologically derived proteins are used in training and inference 4. No hallucinated spaghetti sequences - they fold because they’re based on natural proteins 5. Canonical residue preservation across design specs and antigen selection - no missing cys no abhorrent P or W residues 6. Actual VHH and antibodies (fragment designs and intact) which look super interesting (read natural) 7. Real levers to pull and modify the inference not just temperature and steps (though we have that too 😘). More on this later. Working on a test set to send to a CRO then publish the preprint and code to get feedback. I think people will like this a lot. I love where it’s going.

@davesnetvault Let’s go dude! I really want to see this!

It's happening!!!!! I was recently awarded a grant from Primordia! I will embark into a very exciting journey worth the challenge working on enzymatic nucleotide synthesis using digital microfluidics aligned with open source science philosophy #Primordia, #Glyxon, #DeSci

@SylvainGariel @jasonkeem8652 Reality layer feedback is all the rage!

For anyone building scientific agents on top of stochastic generative tools, another proof that the bottleneck right now is the evaluate-and-filter loop, not the model and not the tool catalog. Striking new benchmark of LLM agents for protein design from @jasonkeem8652, @romerolab1. 76 tasks, 17 tools, four frontier models head to head. Key fact: across 836 runs, agents never once threw out a candidate. They treat each draw from RFdiffusion or ProteinMPNN as the final answer rather than one sample among many to filter. Scoring tools get called at ~14% of expert intensity. Force the agent to generate multiple candidates, score on several metrics, rank, and submit the best, and GPT-5 jumps +15.9 points, DeepSeek V3 +9.3. A control with the same compute but narrower metrics recovers almost nothing. So it is variety of evaluation, not effort. Behavioural gap, not capability. Models are trained to give 1 concise high-probability answer. Protein design rewards generating, comparing, filtering many samples. biorxiv.org/content/10.648…

@MicahOlivas1 @OmicsOmicsBlog I think a qpix is actually cheaper than a facs machine too.

You can do this easily without FACS, just pick single colonies, do colony pcr with per well barcoded primers, deep sequence, then sort. Get fancy and and forward primers have row and reverse primers have column and you have an even lower cost method for screening multiplex libraries.

Many experiments in biology happen one protein at a time, which means synthesizing DNA one gene at a time. This is fine for tens of genes. For thousands, the cost is unsustainable. Introducing uSort-M: a method to isolate and sequence-verify thousands of genes at low cost

Fabricagen’s first summer Intern started today. We have something really fun cooking. He’s a system engineer who placed in a recent academic competition at uni He approached me to advise on that project and we just kept rolling Super smart young man and easy to work with We are going to take a currently available OS protein tool to the next level and rerelease under MIT with new levers to pull Looking forward to sharing our progress later in the summer

The model stack are training signal is really trending nicely Something new is coming Looking around at real lab spaces Someone is working with me now He and I are applying for some funding Intern starts next week too He is working on a cool OS project I hope to share end of summer I haven’t been posting as much because I’m so nose down but it’s going to be worth it

@biocheMichael 🚀🚀🚀🚀

@CotetTudor @benchling Congrats! Great milestone.

I've been using @benchling ever since I first learned to hold a pipette. It has carried me through so many lab classes during uni...so working on this integration has been a full-circle moment