It was great hacking with you all today! #reAgent2026

Genome-scale AI is moving from prediction to programmable biology. On August 10, Ben Fishman from @GenomeLayer will join Physalia for a live webinar on how DNA foundation models can be used to: - predict gene regulation - interpret genetic variants - design functional DNA - connect genomic models directly to agents like Codex, Claude for Science & Phylo We’ll also demo the Genomic Intelligence platform, API + MCP live. If you’re working at the intersection of AI × biology, come join us 🧬 physalia-courses.org/courses… #Genomics #AI #BioTech #AIforScience

Enjoy! Let's know how it works for you! app.genomicintelligence.ai/e…

One GLP-1 response variant became 229 new hypotheses worth testing We analyzed 2,734 DNA changes around GLP1R and found hundreds predicted to have larger effects, including rare variants that population studies may miss This is the opportunity for genomic AI: turning overlooked DNA into testable biology across drugs and diseases genomicintelligence.ai/blog/… #GLP1 #GenomicAI #DrugDiscovery

Very exciting to see Marc Amil turn our genomic models into a working sequence-design loop. When we launched the Agentic Genomics Build Challenge, we invited people to build a practical genomics use case with our API, MCP server, or platform. We promised up to $1,000 for our favorite one (genomicintelligence.ai/blog/…). Marc’s prototype generates candidate sequences, evaluates them with Genomic Intelligence, and iteratively improves them toward a specific biological objective. This is exactly the direction we envisioned: moving beyond predicting what a DNA sequence does toward designing sequences for a specific purpose. Great work, Marc, and thank you for building with us! More challenges to come! What would you build with a genomic model? 🧬

Why does Ozempic work better for some people than others? A Nature study linked a GLP1R variant to greater weight loss. Our genomic models suggest a second mechanism: the variant may affect not only how the receptor works, but also how much of it the body produces. We then screened rare variants that GWAS can miss—collectively carried by an estimated 527M people worldwide. This is what genomic AI makes possible: moving from population-level association to individual, testable biological hypotheses. Read the study and try the workflow on your own genome: genomicintelligence.ai/blog/… #Genomics #GLP1 #PrecisionMedicine #AIforScience

The genome is becoming programmable through natural language We partnered with @PhyloBio to bring @GenomeLayer models into Biomni Lab so researchers can move from a biological question to regulatory DNA analysis, expression prediction, variant interpretation, and sequence optimization inside one agentic workflow Try it yourself ↓ genomicintelligence.ai/blog/… #AIForScience #Genomics #Biotech