IN SILICO
Computation alone
- ×Different targets and datasets
- ×Inconsistent evaluation conditions
- ×High scores do not prove binding
- ×Negative results rarely surface
POREVA OPEN / 2026GLOBAL AI PROTEIN DESIGN CHALLENGE
01 / THE LAB
Every discovery
starts somewhere.
Follow a researcher beyond the screen, into the extraordinary world of proteins.
Explore the challengePO / MOLECULAR EXPLORERCONCEPTUAL WORLD · NOT A MOLECULAR SIMULATION
Preparing your journey…
WHY POREVA OPEN
A computational score is only the beginning.
Teams use different targets, datasets and evaluation methods. A design that scores well in silico may still fail to express, remain stable, bind selectively or produce the intended biological function.
VS
IN VITRO / IN CELL
Real experimental evidence
- Unified tasks and protocols
- Standardized double-blind execution
- Multi-layer functional validation
- Positive and negative data released
THE RULES OF TRUST
One open, credible and consistent measure.
Core rules are published and frozen before submission so different models and methods can be tested under conditions designed for direct comparison.
Rules frozen in advance
Tasks, protocols and the scoring framework are released before submissions open.
FROZEN RULESOne experimental system
Selected designs move through the same standardized experimental workflow.
UNIFIED PROTOCOLDouble-blind execution
Design identity and lab execution are separated to reduce human influence.
DOUBLE-BLINDComplete data matters
Successful and failed outcomes are organized for release where IP and compliance allow.
OPEN RESULTSFROM DESIGN TO EVIDENCE
Scroll through the path from sequence to reality.
Models can generate candidates quickly. Poreva Open moves those digital designs through one traceable experimental pipeline, producing evidence at every stage.
AI protein design
Generate candidate sequences with any model or technical approach.
SEQUENCE / STRUCTURESubmission & freeze
Submit sequences with reproducible method metadata — model version, seeds, templates — sealed at the deadline.
SEALED · SHA-256 + UTCStandardized wet lab
Run scaled expression, purification and binding assays under one protocol.
EXPRESSION / BINDINGAdvanced validation
Top designs enter cellular, physicochemical and organoid evaluation.
FUNCTION / DEVELOPABILITYResults and open data
Publish comparable results and release positive and negative evidence over time.
OPEN BENCHMARKEvidence at this stage · 01Journey progress00%
MGSSHHHHHH
Candidate sequence · Residues 1–10 / 128In silico
Conceptual workflow · not live experimental data
MULTI-LAYER EVALUATION
Binding is one signal, not the whole answer.
The first edition evaluates representative experimental dimensions across the protein design process, measuring generalization, success rate and R&D efficiency of complete design systems.
EXP
Expression
Expression level, yield and fundamental quality.
BND
Binding
Affinity, kinetics and dose response.
SPC
Specificity
Selectivity, off-target and cross-reactivity.
FUN
Cellular function
Functional activity and mechanism in cellular systems.
PHY
Physicochemical
Stability, solubility and related properties.
DEV
Developability
Manufacturing, development and translation potential.
COMPETITION 2026
For everyone serious about solving protein design.
Research groups, AI teams, biotechnology companies, academic labs and independent researchers can participate. No model or technical pathway is prescribed.
TARGET & TASK / 01
The first target and task are coming soon
Challenges will be grounded in real drug discovery needs. Targets, inputs, protocols and the evaluation framework will be published before submissions. Hidden targets and blind tasks will follow in future editions.
Public targetsHidden targetsBlind tasksReal R&D needs
COMPETITION TIMELINE
From frozen rules to open evidence.
Except for the planned November 2026 opening, exact dates will be added after the core rules are finalized.
01
Rules and task release
Publish the target, submission requirements, protocols and evaluation framework.
02
Submissions open
Accept protein designs from teams and researchers worldwide.
03
Unified wet-lab testing
Megalaxy Lab performs standardized double-blind validation.
04
Advanced validation
Top designs move into cellular and advanced biological models.
05
Results and open data
Publish results and progressively release complete experimental data.
WHERE VALIDATION HAPPENS
Megalaxy Lab
All experimental validation is planned for Megalaxy Lab, with scaled protein expression, purification, binding assays, cellular function and organoid capabilities under end-to-end sample and data traceability.
EXPRESSIONPURIFICATIONBINDINGCELL FUNCTIONORGANOID
BE THE FIRST TO KNOW
Let the next sequence speak through experiments.
Registration, the first target, rules and exact dates will be released through Megarobo's official channels.
Read the competition announcementPoreva Open · Let protein design speak through real experiments.