Designed to be scalable and with closed-loop experimentation, Tesseract will power AI to map sequence → function, predict gene transfer success, and assign function to unknown genes from rich phenotypic signatures.

Big thanks to all the collaborators and reviewers who helped shape this proposal!

5/5
We understand this may cause inconvenience, and we appreciate your patience and continued support as we work to deliver a high-quality, impactful Tournament. For any questions or concerns, please contact us at [email protected]

4/5
The new timeline is as follows:

-Registration deadline: Nov. 14th 2025

Predictive phase:
-Zero-shot track: Dec. 1st 2025 to Jan. 16th 2026
-Supervised track: Jan. 26th to Feb. 20th 2026

Generative phase:
-Round-1: Mar. 9th to Apr. 3rd 2026
-Round-2: Jul. 20th to Aug. 14th 2026

3/5
The Tournament will establish the first-of-its-kind benchmark for predictive and generative AI/ML models in engineering PETases. We remain committed to enabling real-world, transparent, and accessible benchmarking competitions, and this adjustment is to uphold that commitment.

2/5
This additional time will allow us to:
•Complete the generation of a high-quality training dataset 🧬🧪
•Secure the best sponsorships to expand access to the Tournament 🖥️💪
•Register more participants 👥🚀

Register before Oct. 17 👉 alignbio.org/get-involved...
More about EvolutionaryScale’s API (Forge) 👉 forge.evolutionaryscale.ai

4/4
What’s coming:
→ More protein functions
→ Sharper ML strategies
→ Public data access portal
→ And yes—we’re hiring.
🔗 alignbio.org/careers
#openbiology #machinelearning #proteinfunction #bioengineering

3/4
GROQ-seq is designed for reproducibility.
Standardized assays + distributed data collection = a foundation for robust, ML-ready protein function data.

Next: scale up function types, refine ML pipelines, expand access.

2/4
🧪 Year One in brief:
Started onboarding 6 protein function types
Launched multi-site runs at NIST + DAMP Lab
Achieved 100K–500K variant throughput per run
Standardized, modular, reproducible, open-source
📘 Full report: doi.org/10.5281/zeno...

🔗 Sign up by Oct 17 → bit.ly/Tournament25
🔗 Want to read more? Last year’s amalyase engineering competition was recently published in PROTEINS: onlinelibrary.wiley.com/doi/10.1002/...
#AIforBiology #ProteinDesign #ClimateTech

4/4
🏆 Why join?:
• $30K in cash prizes
• Free DNA synthesis + wet lab testing
• Leaderboard visibility
• Publication in a special issue of Proteins
• Sponsored conference presentations

3/4
The tournament has two phases:

🔹 *Predictive* (Nov 3 – Dec 23):
Predict PETase sequence properties
🔹 *Generative* (Jan 5 – Apr 10): Design novel sequences to maximize enzyme activity.

Every submission is synthesized and tested.

2/4
This is a *lab-validated* benchmark competition for real protein function.

The winner will be determined by experiment. We’ll test:

🔥 Enzymatic activity at high temperatures
🧬 Protein expression
Results go straight to the leaderboard.

🔗 Register by Oct. 17: bit.ly/Tournament25

Read the news release here: alignbio.org/pressrelease...

🧬 Growth + morphology data across diverse conditions
🔗 Linked to genome references with robust metadata
📊 Built for ML training, benchmarking, reproducibility

🙌 Thx to Jensen Lab for their foundational work that helped enable this platform. Lean more here: alignbio.org/datasets-mic...