Build hybrid talent
Train students to reason with algorithms, biological systems and experimental evidence in one workflow.
About AI-SynBio
AI × SynBio Dry-Wet Innovation
AI-SynBio Challenge is a global interdisciplinary research competition hosted by Tianjin University, championing a dry-wet loop where computation drives experiments and experiments validate models.
Positioning & Mission
Build an interdisciplinary research training system
The competition trains students to reason with algorithms, biological systems and experimental evidence in one workflow.
It moves AI-enabled synthetic biology from promising models to testable, reproducible and useful projects, while connecting students, mentors, researchers and partners through Wiki, Expo, review and open resources.
Dual track, evidence first
AI / Computational dimension
ML/DL models for protein structure, gene-circuit performance or metabolic-flux optimization.
Synthetic biology dimension
Design, assembly and characterization of standardized DNA parts.
Dry-wet closed-loop validation
Competition Orientation
Core Features & Evaluation Directions
Advance frontier applications
Move AI-enabled synthetic biology from promising models to testable, reproducible and useful projects.
Create an international platform
Connect students, mentors, researchers and partners through Wiki, Expo, review and open resources.
Reproducibility
Computational work should record dependency versions, hardware, random seeds, data scripts and model checkpoints.
Experimental traceability
Wet-lab work should provide protocols, controls, experiment notes and verifiable records such as ELN + Git + data fingerprints.
Human Practices
Teams should connect technical decisions with real needs, AI ethics, biosafety, stakeholder feedback and responsible communication.
Dry-Wet Loop
A loop that turns ideas into evidence
AI predicts
A model proposes hypotheses, sequences, targets or experimental priorities.
Wet lab validates
Experiments confirm, falsify or bound the model output with controls.
Data feeds back
Structured results become training, evaluation or calibration data.
Model iterates
The next design is informed by evidence rather than intuition alone.
Organization
Hosted by Tianjin University and supported by specialized committees
Committee
Organizing Committee
Season operations, rule making, dispute handling and cross-team coordination.
Committee
Academic Committee
Scientific direction, strategic positioning and judging-standard guidance.
Committee
Judging Committee
Project scoring, defense review and award recommendations.
Committee
Safety & Ethics Committee
Biosafety, AI ethics, Check-In review and risk assessment.
Committee
Technical Support Committee
Website, Gitee repository, data infrastructure and submission support.
Evidence Requirements
Computational design and experimental validation should support each other
Show how model outputs inform experiments, how results refine the next design, and keep records that can be reviewed.
| Requirement | AI / Computational | Synthetic biology |
|---|---|---|
| Research question | Models propose hypotheses, sequences, targets or experimental priorities. | Experiments confirm, falsify or bound the model output with controls. |
| Process evidence | Code, data scripts, dependencies, seeds and checkpoints are recorded. | Protocols, controls, experiment notes and traceable records are provided. |
| Iteration loop | Structured experiment results become evaluation or calibration data. | Wet-lab outcomes reshape the next computational design. |
| Deliverables | Wiki, repository and reproducible computational methods. | Wiki, protocol, parts or validated experimental evidence. |