Project title
Lignin aromatic biomanufacturingAI-driven value conversion
T3 BiomanufacturingLignin AromaticsHPLC & Enzyme Assays
Much lignin is burned or used at low value despite its renewable aromatic potential. Tianjin University's AroSynth uses graph neural networks for saturation mutagenesis scans, prioritizing candidates for wet-lab tests. Experiments then refine the team's understanding of pathway bottlenecks.
- Team name
- AroSynth_SSBB-TJU
- Institution
- School of Synthetic Biology and Biomanufacturing, Tianjin University
Team media is for event showcases and reporting only. Portrait rights remain with the individuals. For concerns about portrait use, contact the organizers. We will investigate and take appropriate action.
Team Motto
“Decoding aromatic enzymes to unlock lignin's potential ”
Background
Underused Lignin Resources
Large quantities of lignin are burned or treated as low-value waste, wasting resources and adding carbon emissions. Key enzyme efficiency constrains conversion, while lignin depolymerization could supply renewable feedstocks for aromatic chemicals.
Research Approach
Residue Interaction Graphs
The project jointly models protein sequence, 3D structure and mutations using residue interaction graphs. Spatially adjacent residues can jointly affect function. Graph neural networks and protein language models extract structural and sequence features to identify promising mutation sites.
- Modeling Targets
- 7 target enzymes
- Computational Method
- Residue-interaction GNN
- Experimental Validation
- HPLC + enzyme assays
Research Process
Development & Validation
- 01
Stage 1
Ranking & Testing Mutations
Design: scan all single-site substitutions across seven enzymes and rank activity-improvement probabilities and hotspot residues. Build: integrate promising mutants into the chassis chromosome by homologous recombination. Test: quantify products and substrate use by HPLC and measure intracellular/extracellular activity. Learn: compare measured relative activity with predicted probabilities using true-positive rates and correlations, then update the model. - 02
Stage 2
Revisiting Pathway Bottlenecks
Low accumulation of two intermediates suggested that two pathway enzymes could be rate-limiting. Overexpression tests assess these hypotheses and feed into the next round of model optimization. - 03
Stage 3
First Experimental Validation
Enzyme assays supported the AI-predicted mutation sites and showed increased reaction activity in the corresponding mutants.
Project Evidence
Computational & Lab Work
Teamwork
Computational, experimental and candidate review records
Team films
