If a DNA test flagged a risk for my child, I’d ask: how common is this change in people with similar ancestry?
A study of 1,070 Saudi newborns found some variants flagged as disease-causing were more common locally than databases suggested.
https://t.co/7vVjrTQzZH
People are more willing to share genetic data for research when they understand exactly how it'll be used and know they can change their mind.
Trust isn't the soft part of genomic research, it determines whether the data exists at all.
Visit: https://t.co/kwPLKzMYl2
Claude Mythos 5.2 is taking on a new challenge in AI for Biology: predicting entire viral capsids.
Rather than building a new protein-folding model, Claude autonomously orchestrated two specialized models — OpenFold3-preview2 and ESMFold2 — to predict overlapping groups of capsid chains and reconstruct the whole structure using symmetry.
Results: 28 of 34 capsids accurately modeled in Anthropic's internal tests.
The bigger story isn't just protein structure prediction. It's a general-purpose AI agent figuring out how to combine existing scientific tools to solve problems that are difficult to tackle with any single model.
From protein folding to whole-virus architecture, AI agents are opening up new possibilities for biological research and gene therapy.
#AIforScience #AIforBiology #ProteinDesign #GeneTherapy
I looked into Abu Dhabi’s 950,000-genome announcement.
My takeaway: assess the research programme alongside the dataset.
My checklist:
• Relevant data
• Access rules
• Named research partners
• Defined programme
https://t.co/bRQ42WfPog
My @ArabLab notes made perfect sense last week.
Today I’m staring at “send details.”
Details of what, Victor?
My follow-up checklist:
• What we discussed
• What I promised
• Who acts next
• By when
Write those four down before leaving the stand.
Another week another show.
This week we have @ArabLab in Dubai. Just 3 days but should be quite hectic as usual.
I believe that my first show was probably in 1998. It’s gone a long way since then.
🧬 AI Biology — Big Tech Weekly | Oct 19–25
Five tech giants, five fronts in AI biology:
🔬 Google — Autonomous wet labs: https://t.co/thN2wer6dU
🧠 OpenAI — Thousands of agents analyze millions of brain cells: https://t.co/bv5TzkYJtd
💊 OpenAI — Acquires Orchestra Bio: https://t.co/hWEEFgaDO0
🧬 Microsoft — BioEmu-2: https://t.co/EIGCkQ8cOA
🧪 Anthropic — Life sciences research collaboration: https://t.co/wOO1e9RRb8
🧫 NVIDIA — RFD4-Proteina & BioNeMo: https://t.co/N2lvOtrmTQ
The competition is moving beyond models toward biological reasoning, experimental feedback, protein design, and drug development workflows.
BioTender | Bridging AI and Biology
#AIBiology #AIforScience #DrugDiscovery
At #ASHG26, today we recognize Wendy Chung, MD, PhD, whose work continues to bridge scientific discovery and clinical impact.
Her groundbreaking work in genomic medicine has advanced gene discovery, genetic diagnosis, and patient care, helping transform lives around the world.
🧬 Anthropic is inviting life sciences researchers to collaborate directly with its team.
The Guest Investigator Program offers 8–12 weeks of joint research, with access to Claude, AI agents, and support from Anthropic researchers.
🔗 https://t.co/uV0q2dRb9q
Heading to Abu Dhabi’s @FutureHealthAD with Dr Peter Loizou, @mydnabank Chief Medical Officer.
DNA sequencing is advancing. But are we helping people understand their results and make better health decisions?
Looking forward to conversations on longevity, consent and precision medicine.
Clean water 💧 is a right for everyone, wherever they live. You can consider this as my “side hustle”!
I’m more determined to keep pushing for clean water in rural communities.
• Practical water testing
• Operator training
• Reliable supplies
• Ongoing support
The commitment continues…..
BREAKING: Nobel laureate David Baker’s team just published TWO Nature papers in one day!
AI-designed proteases from scratch:
🧬 Cysteine proteases: 19% success rate, functional in mammalian cells.
🧬 Metalloproteases: 36% success rate, with over 10 BILLION-fold catalytic acceleration vs. the uncatalyzed reaction.
From predicting protein structures to designing catalytic function.
📄 Nature papers:
https://t.co/tURwbve5kT
https://t.co/igz434xxv0
Nextflow 26.10 is here! 🚀 Static typing is now stable, and a year of language work comes together to make your pipelines composable, reusable building blocks that you (or your AI agent) can verify before a single task runs.🧵https://t.co/Uf54cZbzoi
Clinical trial recruitment came up repeatedly in my conversations in Basel.
Hospitals overestimate eligible patients. Records are fragmented. Screening is slow.
Pharma spends millions trying to find patients whose data often already exists.
Is the real problem recruitment or access to reliable data?
https://t.co/PNEd6T2xhE
Introducing Agentic Nextflow: AI agents as steps inside your pipeline 🤖 Agent steps get the same guarantees as every other step: reproducible, portable and scalable. One engine for deterministic and agentic compute.📗Read the blog: https://t.co/iygeJ8XODs #NextflowSummit
Two children avoided invasive procedures after rapid genomic testing identified non-cancerous conditions.
In a 54-child study, reports averaged 3 days versus 42.
As a parent, faster answers matter.
What delays results most?
https://t.co/qNcV7y5yCP