[Showcase] Autonomous generation of a 5°C marine PET-hydrolase

[n57d30top]

(Posting this as a showcase issue since the Discussions tab appears to be disabled. Feel free to close or convert if needed!)

Hi everyone,

We wanted to share a successful downstream application of autonomous protein design with the community. While current PET-depolymerization bottlenecks heavily around the thermodynamic requirements of mesophilic/thermophilic enzymes (e.g., FAST-PETase needing 50°C+), we tasked our internal cluster (Sovryn Oceanus) to search for a cold-saline adapted candidate.

(To clarify upfront: This sequence was designed end-to-end entirely by our proprietary architecture, independent of OpenFold or AlphaFold prediction weights. We are sharing it here strictly to engage with the world's best structural biology community.)

The engine autonomously generated oceanus_petase_variant_3, which demonstrates stable predicted folding and catalytic geometry specifically calibrated for ester hydrolysis at 5°C in marine salinity profiles (Catalytic Proxy Score: 0.756). This provides a theoretical baseline for passive, unheated oceanic micro-plastic remediation.

To prevent commercial enclosure before synthesis is verified, we anchored the .pdb structural coordinates and the sequence to the Ethereum Base L2 blockchain (TxID: 0x3ec895c0b9ce325944a40f595a3859dd9e3e605841190c63d11dda27120c3f10).

GitHub Repo (including the raw PDB & L2 Hash): https://github.com/n57d30top/oceanus-pet-hydrolase-5C/tree/main

We have released the architectural parameters and the explicit PDB file under a Dual-License (free for academic/NGO remediation, restricted for commercial use).

We are currently looking for BSL-authorized wet-lab partners to transition this from in-silico desk review to physical synthesis and falsification. Any feedback on the predicted psychrophilic adaptations from the computational community would be highly appreciated.