Really impressive work just published in Nature Machine Intelligence by the @NREL + @harvardmed teams. They trained 11K+ models to predict enzyme pHopt — and PLMs dominated across the board. Biology meets language models → 🔥.

Open benchmarks move the field forward. The 2024 Adaptyv EGFR binder comp just dropped its preprint: Real target, real validation, public data. Great step for reproducibility + collaboration in protein design. #ProteinDesign #Adaptyv #ML #OpenScience.

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

Excited to build alongside this next phase. Link: @UWproteindesign . #rfdiffusion2 #proteinengineering #AIenzymes.

RFDiffusion2 is here — major upgrade for active site conditioning in enzyme design 🧬. Define key residues + ligand → get much better scaffolds. Same pipeline, stronger context. Plenty of momentum:.• Activity’s improving.• Hit rates rising.• Scaling’s tough — but solvable.

Need a better enzyme?. We help life sciences teams get custom, industrial-grade enzymes — fast. Struggling with: .🧪 Low selectivity.🧬 Poor scalability.🕒 Long optimization cycles?. We’ve got you. Learn more → #AI #Biotech #EnzymeDesign.

@VanderbiltU The future of protein design isn't just AI-driven — it's AI-informed, physics-aware, and purpose-built. 🔗 #ProteinDesign #AIinBiotech #SyntheticBiology #ProteinEvolution #ComputationalBiology #Biophysics #EnzymeEngineering.

@VanderbiltU By pairing state-of-the-art structure and sequence prediction tools with thermodynamic and functional constraints, we're making it faster, cheaper, and more intuitive to design proteins that solve real-world problems — from polymer degradation to therapeutic development.

@VanderbiltU This integrated approach closes the gap between theoretical design and real-world function — a challenge every protein engineer knows well. 🧬 At Protein Evolution, we’re seeing the same value in hybrid models.

@VanderbiltU The team demonstrated that AI alone can generate novel protein sequences, but by layering in physics-based folding models, they can validate and optimize those sequences with much higher confidence.

Recent research from Leipzig University and @VanderbiltU shows how combining artificial intelligence with biophysical modeling significantly improves our ability to predict how proteins fold — and importantly, how they can be redesigned for new functions.

While these findings are early-stage, they suggest exciting possibilities for biotech, pharmaceuticals, and novel chemistries. #AI #Biotechnology #EnzymeEngineering #DeepLearning . Read more here:

CataPro leverages pre-trained models and molecular fingerprints, showing potential improvements over existing methods, which may significantly enhance enzyme discovery and optimization.

However, a recent study published in Nature Communications introduces CataPro, a promising deep learning model designed to predict enzyme kinetic parameters (kcat, Km, and catalytic efficiency).

🚀 Cautious optimism for AI-driven enzyme discovery!. Historically, AI models have had difficulty accurately predicting enzyme behavior due to the complexity of relating enzyme structure to catalytic performance.

@arcinstitute That’s why we focus on combining these innovations with human-in-the-loop refinement — to reduce experimental cycles and make discovery more efficient. We’re eager to see how tools like this continue to push the field forward. #ProteinEngineering #GenerativeAI #SyntheticBiology.

@arcinstitute At @ProteinEvo, we’re constantly learning from advancements like this. AI-driven tools are reshaping protein engineering, but translating designs into real-world results still requires careful iteration and expert insight.

Exciting progress from the @arcinstitute! Their latest generative AI tool marks a significant step forward in protein design, helping to accelerate discovery across the life sciences, human health and longevity, and environmental innovation. 🔗

At Protein Evolution, we’re committed to changing this. Our mission is simple: make plastic from waste, not oil. We've engineered enzymes that break down complex polyester at the molecular level, enabling true circularity and significantly reducing greenhouse gas emissions.

When we think of plastic waste, most of us picture bottles and bags — but did you know that textiles are an even bigger issue?. With fast fashion on the rise, the problem is only growing.

Circularity may be a buzz word – but for us, it’s a technical term. Our Biopure process achieves true circularity with plastic production and end-of-life waste.