Derya Unutmaz, MD ✓

[image] It is now conceivable that we might achieve biosingularity much earlier than I had predicted 20 years ago (circa 2040s). I had coined this term for a time when we will be able to completely engineer and reengineer biological systems and even create new organisms and functions.

[image] The best way to follow BioAI news is via Peter’s BAIO newsletter—so much is happening that it’s becoming increasingly difficult to keep up!

[image] This could be the current peak of biohacking! I believe I was one of the first to use the word Biohacker in the mid-90s. Sequencing your own genome wasn’t even a dream then! Years after that, it cost $3 billion to sequence the first genome! Now you can have it for few hundred $!

[text] Wow! One of the most exciting BioAI news announcements of the year! From @GoogleDeepMind in collaboration with @emblebi, @Nvidia and @SeoulNatlUni ! “Millions of protein complexes added to AlphaFold Database. Four-way collaboration brings together world-leading AI and

[image] I agree with this. We need tons more biological data and much more compute to analyze & simulate biological systems with AI. This is why my timeline for curing all diseases is around a decade & reversing aging is in 15 years-most of that time is for getting the data & the compute

[image] Biomni is one of the top BioAI models, which is very useful for biomedical researchers. I highly recommend it! Also, congratulations to Kexin and the team for the Science publication!

[video] A new BioAI era is beginning. It is now possible to have your AI agents directly test their protein designs experimentally in the wet lab through the Adaptyv API!

[video] This is so exciting! BioAI just jumped a level with the release of ESMFold2! Kudos to Alex and the team for making this open scientific engine and releasing billions of proteins and predicted structures!

[text] Anthropic is hiring scientists with biology backgrounds. This could be a great opportunity for those interested in BioAI.

[video] Super excited to test Biomni-Tuso over the weekend! I was a very early tester of Biomni & believe it has now become the top BioAI platform available. Recommended it to several biomedical scientist friends and they all love it! Biomni-Tuso is looking to push the envelope further!

[image] This is a fantastic advance in genome-wide editing with CRISPR! Neville’s team has developed amazing tools as CRISPR fusion proteins! Congrats, guys!

[video] Great BioAI model from Insilico Medicine! “By connecting PandaOmics directly to your favorite AI agents and tools, users can access target discovery, pathway analysis, and biomedical intelligence seamlessly.”

[video] I'm really excited about this! A top open-source BioAI model can make an incredible impact in finding new treatments! I'll need to try it out soon!

[image] PandaClaw is the coolest agentic AI toolset for BioAI! Very excited about this massive release of 140+ scientific skills and 1,000+ bioinformatics tools! This is great, Insilico Medicine! Let’s have AI🦞 help decoding biology!

[image] I'm super excited about this agentic BioAI model from NVIDIA! We have now entered a new era in science, which can be termed Science 3.0! In 1.0, individual scientists made most of the contributions; in 2.0, it was teams and big data generation; and in 3.0, it’s AI on top!

[image] One of the best BioAI models, Evo 2, is now published in Nature.

[video] Amazing BioAI resource! From Nvidia in collaboration with @GoogleDeepMind @emblebi and @SeoulNatlUni

[text] This is a great BioAI model! It also demonstrates how vast biological data is: 27 terabytes of tokens just for GI endoscopy data! Eventually, we will need to train BioAI foundation models with petabytes of data, thus we need much more compute! This will save so many lives!

[video] BioAI is getting agentified! Operon agent would be super cool to have!

[image] Great news from NVIDIA for BioAI builders! AI is democratizing genome analysis!

[image] Great BioAI review!

[video] This is a super cool science and BioAI platform! Can’t wait to try it out! Indeed, as I’ve been saying for the past two years, we have now entered a new era in Science, which I call Science 3.0, where AI and human scientists work together to push the frontiers of discovery!

[image] Amazing advance by Qwen3.8-Max on BioAI reasoning, benchmarks tested by @InSilicoMeds

[video] AI-designed 3D ligands that bind perfectly to target pockets are the holy grail of AI-based drug design! We're getting very close; soon AI will design molecules better than available drugs. People still don't understand how amazing AI's impact will be for human health.

[image] It's interesting that Grok 4.5 from @SpaceXAI tops the longevity benchmark. Let's see if Grok 4.6 can compete with GPT-5.6 Sol and Opus 5 in other drug discovery benchmarks.

[image] The dream of an AI biologist is now becoming a reality! Xunzi sounds like a superb BioAI model. Already forked the source code:

[image] All I can say is this is an amazing BioAI tool! Current revenue for all peptide therapeutics combined is estimated at over USD $100 billion! Better peptides could greatly reduce the cost of these drugs and also expand the market through much greater use!

[video] I don’t have access to GPT-Rosalind, but it looks amazing! I’ve worked with all the proteins they mention in the video, so it resonated with me a lot. I hope I can get access to test it at some point, for now we will drool 😅

[image] Soon it will be difficult to imagine doing science without AI, both as collaborators and as tools! BioAI will have the highest impact because nothing is more complex than biological systems, and I’ve been saying that decoding them would have been nearly impossible without AI!

[image] This is so cool and beautifully made! 😲 Love it!

[video] My friend @biogerontology explains why it is important to open AI benchmarking tools to entire drug discovery industry: 🔹 Open benchmarks to help more models get smarter 🔹 Feed real drug R&D data back into AI 🔹 Turn breakthroughs into out‑licensable assets “At the end of the

[image] Insilico Medicine DDD Benchmark is now the best clinical drug discovery platform! Big congrats to the awesome team! @AlexAliper @biogerontology “Huge Drug Discovery Foundations: 300+ evaluations across disease biology, property prediction, retrosynthesis, structure-based

[text] Very excited about ProteinDPO because it closes the critical “alignment gap,” enabling GenAI or BioAI models to optimize for real experimental properties like protein stability instead of just producing plausible sequences. This will in turn make AI-designed proteins far more

[image] Amazing new virtual tissue foundation model has just been published in Nature! “Virtual Tissues (VirTues), a general-purpose foundation model for spatial proteomics that learns marker-aware, multi-scale representations of proteins, cells, niches and tissues directly from

[text] I started this community in June 2024, initially to follow the remarkable advances emerging from the release of AlphaFold 3. Since then, the community has grown to more than 3,300 members, reflecting the extraordinary expansion of interest in AI across biology, biotechnology,

[image] This benchmark issue is really important to advance BioAI. Alex explains this well in this thread.

[image] Excellent BioAI benchmark! We need more of these to test new models for biological systems.

[image] Very interesting BioAI model for the gut microbiome (an area of great interest for me as well). The code is open source. I hope to check it out soon!

[image] Another super exciting BioAI model, BioReason Pro, released by Bo and his team in collaboration with @arcinstitute and @genophoria

[image] Major SynBio AI model just dropped! : 1. For human health, Omnii can read and write whole genomes (more on writing later). It’s state of the art (SOTA) on detecting causal variants for disease, and can rank Alzheimer's mutations zero-shot. We’re partnering with a diagnostics

[text] Great database of BioAI models from Michaela. See the GitHub repos.

[video] Amazing whole body BioAI foundation model from Ali’s lab! Congratulations for the Nature paper!

[image] Very cool new BioAI model called dnaHNet from the Arc Institute!

[image] I've been testing LiteFold. It's an excellent BioAI model!

[image] These AI-designed antibody drugs by Nabla Bio look awesome! I am particularly excited about the antibodies to MHC-peptide and TCRs. I created the infographics below to explain what this means and why it is important.

[image] New BioAI multimodal model DISCO simultaneously designs protein sequence & 3D structure from scratch!

[video] Development of virtual cells is the boss-level challenge of BioAI! Bo and the team at Xaira seem to be making great progress in this!