The Profluent team is heading to Summer RosettaCon next week! Find @jeffruffolo, @richardwshuai, @ShiakiMinami, and Alex Hoffnagle and say hi 👋 You can hear Richard talk about E1, our encoder model, on Thur at 1:30pm (“E1: Retrieval-Augmented Protein Encoders for Fitness and Structure Prediction”) and Alex share his work on our AI-designed base editors on Mon at 7:30pm (“Design of Programmable Base Editors with Protein Language Models”).

Read the PR here: prnewswire.com/news-releases…

A suite of AI-designed base editors, all from one scaffold. The only piece unique to each patient is the guide RNA. That's the shift AI unlocks: rare disease moves from 'one drug at a time' to a platform, built to bring cost down and reach more patients. More on our work with GEMMABio, now backed by @ARPA_H's THRIVE program, below.

The bottleneck in rare disease is design: too many mutations, one editor at a time. AI unlocks that. Proud to be building GemmaBio's platform with them under @ARPA_H 's THRIVE program. We're bringing our frontier AI models and base editors to the work to design modular gene editors that scale across many mutations.

P.S. Yes, we are hiring! job-boards.greenhouse.io/pro…

And here's a few places you can find us during the conference: — Oral talk: @thisismadani with collaborators from NVIDIA and Microsoft Research: "FLIP2: Expanding Protein Fitness Landscape Benchmarks for Real-World Machine Learning Applications" (Tue, Jul 7 @ 10:15am, Hall D2) — Workshop: Join @jeffruffolo, @AadyotB, and the Profluent team: "E1: Retrieval-Augmented Protein Encoder Models" (Sat, Jul 11 @ 2:25pm, Room S317) — And our very own @park_jungy will be presenting work from his PhD: "Discovering Symmetry Groups with Flow Matching" (Wed, Jul 8 @ 5pm, Hall A #3010) and "Smoothness Errors in Dynamics Models and How to Avoid Them" (Thu, Jul 9 @ 2:30pm, Hall A #2200)

The Profluent team will be at ICML in Seoul next week. Find us to chat, grab a coffee, or join us for dinner! We’d love to talk shop (protein language models, protein optimization, sequence-first methods vs structure-first). Basically, anything at the intersection of AI and biology. Want to grab coffee? Let us know here: forms.gle/yWQpAvLiQ1KfM85u7 Or want to hang out and chat over dinner? Reserve a spot at our table: luma.com/qpc1mfw9

We're still in the GPT 1.5 era of AI × biology The early models already work (signal: our $2.25B Lilly partnership) but we're nowhere near the ceiling We're speedrunning to GPT 5 as fast as we can @thisismadani with @nathanbenaich @airstreet

Biology is no less complex than text and (in our biased opinion) way more impactful. Yet a fraction of the world's AI talent and compute is pointed at it. The field is wildly undersaturated. Watch @thisismadani chat with @nathanbenaich @airstreetcapital about the opportunity.

OpenCRISPR was the first demonstration that AI could design a genome editor from scratch. We used our AI model to build a protein that doesn't exist in nature for a specific function and it worked. Watch @thisismadani chat with @nathanbenaich @airstreetcapital about where we’ve gone from there (and where we’re going)

Our partnership with Eli Lilly carries up to $2.25B in milestones. The bigger story is the unlock behind it: large gene insertion, a problem AI makes solvable for the first time. Watch @thisismadani in conversation with @nathanbenaich @airstreet.

At PEGS Boston? Don't miss our Lead Protein Design Scientist Jeliazko Jeliazkov presenting "Designing Optimal Proteins at Scale" Generating proteins that are simultaneously optimal across many properties (affinity, stability, developability, and beyond) is a hard problem. Jeli's sharing our work on alignment of our foundational AI models as a path to multi-parameter protein optimization, with applications from gene editors to antibodies. Interested in learning more about our multi-parameter optimization work? Shoot us a DM.

We're at @ASGCTherapy today sharing something we've been heads down building: using our AI models to scale base editing for personalized medicine. The gap between what's theoretically correctable and what we can actually fix today is huge. We think AI can close that gap. Not there? Peter Cameron, our SVP of Gene Editing, breaks it down here. Interested in learning more? Shoot us a DM.

Our frontier AI models design custom recombinases from scratch, programmable to target virtually any location in the genome. We're collaborating with @EliLillyandCo to turn that capability into medicines. Read the press release for more: businesswire.com/news/home/2…

This has been a long sought goal in the gene editing field, but current tools can't reliably make insertions at that scale. Naturally occurring recombinases could but are limited in where they can act and traditional metagenomic discovery and protein engineering approaches can't precisely control their targeting.

Many genetic diseases involve hundreds of different mutations across patients–no single edit can address them all. Inserting a complete functional gene with a recombinase editor opens the potential to address diseases with high mutational heterogeneity using a single therapeutic, rather than developing a separate editor for every variant.

Today we announced a landmark partnership with @EliLillyandCo to use our AI models to design recombinases for genetic medicine—a collaboration valued at up to $2.25 billion before royalties. The goal: use Profluent's AI models to design recombinase editors capable of inserting long stretches of DNA at precise locations in the genome. Read the press release for more: businesswire.com/news/home/2…

What do AI-designed proteins look like in the field? 🌾🚜 On Monday, @thisismadani joins @Corteva at @WorldAgriTech to share how Profluent's AI models are engineering proteins for gene editing solutions that address real problems farmers face. March 16 · 12:30pm · San Francisco · World Agri-Tech worldagritechusa.com/agenda-…

📄 Read the full paper here: nature.com/articles/s41587-0… 🧑‍💻 And explore the tool: protein2pam.profluent.bio

We’re excited to share our latest work published today in @NatureBiotech: Protein2PAM, an AI model that enables the rapid design of CRISPR editors with new PAM recognition And we’re making the model freely available for research and commercial use: protein2pam.profluent.bio