Johns Hopkins UniversityEst. fall 2025~30 members

UPET: Undergraduate Protein Engineering Team

No experience required.

Competition teams
Anthropic × Adaptyv 2026 · Track 1 · CASP17 · PETase 2025
Research teams
Original protein engineering research
Top7 · PDB 1QYS
Sequence · residues 3–94

Amino-acid sequence of Top7, residues 3 to 94: DIQVQVNIDDNGKNFDYTYTVTTESELQKVLNELMDYIKKQGAKRVRISITARTKKEAEKFAAILIKVFAELGYNDINVTFDGDTVTVEGQL

01Track 1Fall 2026

Anthropic × Adaptyv

Protein Design Competition

UPET · one of ~30 teams worldwide · selected by Anthropic

  • $50,000in Claude credits
  • $10,000in Modal compute
  • 100designs tested in Adaptyv’s lab

Selected forTrack 1

Track 1 · for labs & companies with protein design expertiseOfficial terms

We’re undergraduates.

Five challenges · one a week, two weeks each · Sep 28 – Nov 8, 2026

  1. 01EGFRBinds at pH 6.5 (tumor) · not at 7.4Sep 28 – Oct 7
  2. 02TNF-αMouse too · binds at pH 7.4, lets go at 6.0Oct 5 – 18
  3. 03GLP1RAgonist · has to switch the receptor onOct 12 – 25
  4. 04TfR1Mouse too · lets go in the endosome (pH 6.0)Oct 19 – Nov 1
  5. 05TBAAnnounced when it opensOct 26 – Nov 8

Wet-lab validation · by Dec 7, 2026Open results · early Jan 2027Competition page

02TeamsCompetition · research

Two kinds of teams.

Competition teams

  • 2026Anthropic × AdaptyvTrack 1 · Track 3
  • 2026CASP17blind structure prediction
  • 2025PETase Tournamentzero-shot · generative design

Research teams

  • NowOriginal researchprotein engineering projects
  • NextPatented designscollaborations with Hopkins labs

The work

  1. AdaptImprove a natural proteinPETase · plastic-degrading enzyme
  2. BindDesign a binder for a targetEGFR · on at pH 6.5, off at 7.4
  3. InventGenerate de novoTop7 · novel fold, 2003
03How UPET worksNo gate

How UPET works

“You can do real, good work, and you don’t have to be screened by a lab or be mentored for years.”

Jonathan Ouyang · founder & president
  1. 01

    No prerequisites

    Any major · any year

  2. 02

    One problem a week

    New challenge every Monday · two weeks to design and submit

  3. 03

    Taught as we go

    Education first · cognitive science to biophysics

04AIWith support from Anthropic

Agents in the loop.

Beyond prediction models · agentic AI across the design loop

  1. 01Reasonover biological data
  2. 02Proposethe next change to the molecule
  3. 03Testin the wet lab
  4. 04Interpretthe results, then again
05RecordFall 2025 → next

One year in.

  1. Fall 2025FoundedJonathan Ouyang, founder · Prof. Jeffrey Gray, mentor
  2. Fall 2025PETase Protein Engineering TournamentZero-shot prediction · generative design · plastic-degrading enzyme
  3. 2025–26CASP17 · PETase · therapeuticsBlind structure prediction · enzyme designs · therapeutic designs
  4. Sep 2026Entered the Anthropic × Adaptyv Protein Design CompetitionSpecialist team · general team (Track 3)
  5. Sep 27, 2026Selected for Track 1~30 teams worldwide · $50,000 Claude · $10,000 Modal · wet-lab validation
  6. Sep 28, 2026Challenge 01: EGFRpH-selective binder · first of five weekly challenges
  7. Oct 2026Challenges 02–04: TNF-α, GLP1R, TfR1Cross-species conditional binders · a receptor agonist · a brain-shuttle binder
  8. Oct 2026Verified under Anthropic’s Life Sciences Verification ProgramGranted to UPET as an organization · vetted for work with Anthropic’s models
  9. Oct 2026~30 membersCompetition teams · research teams · growing
06People~30 members · growing

Who we are.

Faculty mentor

Prof. Jeffrey J. Gray

Professor, Chemical & Biomolecular Engineering · Whiting School of Engineering

Gray Lab · structure prediction & design · PyRosetta · Rosetta

~30 members · majors

  • Biophysics
  • Biomedical Engineering
  • Chemical & Biomolecular Engineering
  • Computer Science
  • Applied Mathematics
  • Cognitive Science
—PressUPET in the news

In the press.

07JoinOpen to every Hopkins undergrad

Join UPET.

Who
Any Johns Hopkins undergrad · freshman to senior · any major
Experience
None needed
Teams
Competition · research