Our project has achieved systematic results in Dry Lab Modeling, Parts, Wet Lab, and Human Practices, constructing a reusable antifreeze engineered yeast chassis. We introduced AFPs gene components, trehalose synthesis gene components, promoter and expression regulation elements, as well as the Cre/loxP site-specific recombination system and the Rci/sfxa101 site-specific inversion system, offering the iGEM community clearly structured, functional, and expandable biological parts and engineering strategies.

1. We developed a computational platform for AFP modeling, integrating machine learning, neural networks, and large language models to help future teams optimize or design new proteins.

2. We designed and constructed two new composite parts: trehalose synthesis and surface-displayed AFPs, enriching the iGEM Parts Library.

3. We established a secretory protein model, from professor discussions to policy alignment, and from topic selection to commercialization in fashion and cosmetics, providing a complete human practice pathway from need to application.

This project has established a systematic and multidimensional computational modeling platform for AFPs. Based on rational design, mutation prediction, and inverse folding methodologies, we employed machine learning, neural networks, and large language models to predict and optimize the structure of AFP candidate proteins. The results obtained through our predictive models and molecular dynamics simulations are consistent with those from wet experiments, demonstrating the reliability and effectiveness of the computational approaches used.

Figure 1. Model Framework Diagram proteins

1. Protein Design
We implemented a three-track parallel computational design strategy, integrating point mutation optimization, inverse folding based on natural templates, and de novo design guided by physical rules to systematically explore the design space of antifreeze proteins. This approach offers future iGEM teams a multidimensional and systematic framework for protein design, enabling comprehensive exploration from functional enhancement to structural innovation.

2. Prediction Model
We developed the AFP-igemTJ2025 model based on the ProtT5-XL language model and a custom deep learning classifier to enable high-accuracy functional prediction of antifreeze proteins directly from sequence data. This model performs exceptionally well on imbalanced datasets and provides a reliable tool for future teams to rapidly screen and annotate AFP candidates.

3. Molecular Dynamics Simulation
Molecular dynamics simulations were used to evaluate the initial dynamic properties of candidate proteins, providing data support for subsequent wet experimental validation. This method allows teams to predict protein stability and functional performance before experiments, improving design efficiency and reducing experimental resource waste.

4. Short Peptide Synthesis
Short peptide design was supported by a comprehensive computational workflow: the k-mer algorithm was used to calculate Jaccard similarity for fast preliminary screening; Bio.Align performed global alignment to improve accuracy; the CKSAAP algorithm analyzed 8-mer spaced amino acid pair frequencies to filter out non-AFP sequences, while a lightweight model based on variance thresholds was developed for efficient AFP identification; HHrepID-inspired methods detected repetitive motifs with high self-similarity; MEGA was used to construct phylogenetic trees, helping identify evolutionarily related protein groups; and WebLogo visualized representative motifs, providing a theoretical basis for further experimental validation.

5. Chatbot
We built a Chatbot on the Coze platform integrated with an AI agent and customized system prompts to accurately answer questions related to our project. Deployed on the WeChat official account, this Chatbot serves as an effective communication and outreach tool for iGEM teams, enhancing the efficiency and accuracy of project dissemination.

The five aforementioned modules collectively constitute a complete dry-lab modeling system for our project, providing future iGEM teams with a reusable computational toolchain that supports the efficient development and rapid validation of antifreeze protein engineering, antifreeze system design, and multi-protein co-expression strategies.

In this project, we have constructed and submitted a series of high-quality standard biological parts (BioBrick), covering topics such as the S. cerevisiae surface display system and the trehalose synthesis pathway. These parts provide the iGEM Village and future teams with well-structured and functionally defined biological components.

We introduced an enhanced endogenous trehalose synthesis pathway in yeast, which includes five key genes, and ligated it with the yeast shuttle vector pRS413 to obtain the plasmid pRS413-trehalose. We then uploaded the new composite parts, enabling the synergistic effect between AFPs and trehalose. These components offer other teams a reusable metabolic regulation module for low-temperature yeast engineering.

Figure 2. Trehalose Pathway proteins

We have introduced 11 naturally occurring antifreeze proteins into the iGEM Parts Registry. In addition, we designed and constructed four point-mutated antifreeze proteins, four domain-fused mutants, four inverse-folding mutants, six de novo designed antifreeze proteins, and four mutants with substituted helical structures. Several of these proteins exhibit superior antifreeze activity. All have been submitted as Basic Parts, offering future iGEM teams working in antifreeze-related fields a comprehensive set of standardized biological components for reference and utilization.

We introduced a surface display system in Saccharomyces cerevisiae[1]. The AFPs were fused with Aga2p via a GS linker to form fusion proteins, which were then expressed in the EBY100 yeast strain, enabling the surface display of AFPs on yeast cells.

To validate the performance of the surface display system, we employed fluorescent labeling to confirm that the target proteins were successfully displayed on the yeast surface. Subsequently, we constructed surface display systems for protein variants MUT4 and NEW11, which exhibit improved antifreeze properties. The effectiveness of the surface display system in enhancing the antifreeze activity of AFPs was successfully verified through IRI assays and yeast freezing-thawing survival rate experiments. These systems provide valuable metabolic regulatory components for future iGEM teams working on antifreeze yeast engineering.

For more detailed information, please refer to Parts page.

Figure 3. Surface Display System proteins

All the BioBrick parts we submitted comply with the iGEM standard format, including annotations and validation data. We invite other teams to access and utilize these parts on the Parts page.

1. Chassis Strain Construction
We constructed an efficient, stable, and scalable antifreeze yeast chassis centered on the integration of the surface display system and the synergistic mechanism of trehalose. This work provides a reference strategy and practical tools for future iGEM teams in developing cold-adaptive chassis strains, and contributes reusable genetic modules and standardized procedures to the iGEM Village.

（1） Surface Display System
We introduced a S. cerevisiae surface display system by selecting the EBY100 strain and the pYD1 plasmid to construct the Aga2p-GS linker-AFP fusion protein. This enabled the surface display of AFPs, thereby achieving antifreeze effects both on the yeast surface and within the cell. The system was validated through IRI experiments and yeast freezing-thawing survival rate assays, successfully producing yeast chassis strains with notable antifreeze activity.

This approach offers a feasible solution to the high cost and low purity issues in large-scale protein purification in engineering contexts, and provides a reference framework for future iGEM teams aiming to construct systems with coordinated extracellular and intracellular protein functions.

Figure 4. Characterization of antifreeze performance in yeast strains AFP1-11, pYD1, and GFP. (a) IRI activity assay. (b) IRI activity assay – mean ice crystal size analysis. (c) Yeast freezing survival rate test at -80 °C. (d) Strain survival rate after freeze-thaw cycles.

Only the experimental results are presented here. Detailed results and analysis can be found on the Wet Lab - Results page.

（2） Trehalose Synergistic Mechanism
We introduced an enhanced trehalose synthesis pathway to improve the survival rate and stability of the chassis strain under extreme low temperatures. The production level was validated by mass spectrometry, and the synergistic antifreeze activity of trehalose with AFPs was confirmed through Ice Recrystallization Inhibition (IRI) assays and yeast freezing-thawing survival rate experiments.

This work provides future iGEM teams with insights into intracellular and extracellular synergistic antifreeze strategies, and offers a portable genetic circuit for constructing multifunctional and stress-resistant chassis strains.

Figure 5. Characterization of the antifreeze performance of the yeast strains EBY100-pR_DK05/09/10/11, EBY100-pRS413-trehalose, and EBY100-pRS413-trehalose-pR_DK05/09/10/11. (a) IRI activity assay. (b) IRI activity assay – ice crystal mean size analysis.

Only the experimental result figures are presented here. Detailed results and analysis can be found on the Wet Lab - Results page.

（3） Biocompatibility
After obtaining antifreeze proteins with favorable IRI experimental data, we further introduced a systematic cytotoxicity assay to evaluate their biocompatibility.

Using L929 cells as the model system, we validated the low toxicity and good biocompatibility of representative antifreeze proteins such as AFP4, AFP8, AFP9, and MUT4. This experiment provides essential foundational data for the application of our project in the cosmetics industry, and also offers an optimization direction for the engineering design of yeast chassis strains.

The introduction of the cytotoxicity test, along with the promising experimental results, has significantly enhanced the credibility of our project. Additionally, we have provided a standardized experimental protocol for L929 cell-based cytotoxicity testing, which can serve as a reference for future iGEM teams requiring biocompatibility assessments.

Figure 6. L929 Cell Cytotoxicity Assay Figures

Only the experimental result figures are presented here. Detailed results and analysis can be found on the Wet Lab - Results page.

（4） Integration of Experimental Methods
We systematically integrated key methods for characterizing antifreeze performance, including Ice Recrystallization Inhibition (IRI) assays and yeast freezing-thawing survival rate experiments, along with a full set of procedures for cytotoxicity testing. This has resulted in a comprehensive and reproducible experimental workflow. This integration provides future iGEM teams working on cold-adaptive strains or aiming to characterize low-temperature stress resistance with a reference for standardized experimental procedures.

For detailed information on the integrated experimental methods, please refer to the Wet Lab - Protocol page.

2. Site-Specific Recombination – Multimer Inversion Combinatorial System
We implemented a combinatorial expression system by integrating the Cre/loxP site-specific recombination tool[2] with the Rci/sfxa101 multiple inversion system[3].The Cre recombinase (tyrosine recombinase) and loxPsym sites were used to enable random shuffling of the AFP expression cassettes.The Rci recombinase (DNA inversion enzyme) and sfxa101 sites were employed to induce promoter inversion, thereby modulating gene expression levels.

By combining these two systems, we screened for the optimal AFP combinations and ideal expression ratios, achieving precise control over the expression levels of the target proteins.

Figure 7. Cre/loxP System for AFP Gene Rearrangement

Figure 8. Rci/sfxa101 System for Promoter Inversion

We have preliminarily constructed the rearrangement plasmid pTJ25-CL and the multiple inversion plasmid pTJ25-MI, and are currently conducting experimental validation. This combinatorial system helps to establish a complete AFP expression library, providing a high-throughput screening strategy for multi-protein synergistic expression. It also offers a comprehensive framework for future iGEM teams aiming to regulate the coordinated expression of multiple proteins.

Through a systematic Human Practice approach, we integrated the modules of Expert, Society, Education, Industry, and Government, clarifying the internal logical relationships among them, and abstracted our workflow and achievements into the "secretory protein" model.

Figure 9. Human Practice Secreted Protein Model

1. Society
The Society module served as our channel for engaging with the broader community, enabling the acquisition of inspiration from public needs, analogous to the ribosome in the secretory protein pathway—functioning as the foundational component of the entire process.

Through multidimensional practices such as the Hibiscus Flower Festival, science communication activities, AI model construction, and questionnaire surveys, we not only enhanced the project’s social impact but also identified key directions for technological application based on real feedback. Additionally, by incorporating culturally representative interactive forms, such as the “Nine-Nines Cold-Weather Chart,” we successfully attracted public interest in synthetic biology and antifreeze proteins.

This module provides future iGEM teams with innovative models and references for human practice, demonstrating how cultural elements and AI tools can be effectively utilized to increase public participation and improve the efficiency of project communication.

2. Education
The Education module acted as a bridge between the project and society, serving as the "rough endoplasmic reticulum" in our model, responsible for the initial processing and dissemination of project concepts.

We designed the "Cell Factory" card game, which transformed the complex process of protein synthesis into an intuitive and engaging educational tool, effectively stimulating teenagers' interest and understanding of synthetic biology.

This section offers future iGEM teams guidance and reference for designing multi-level educational activities and establishing mechanisms for feedback collection in human practices.

3. Expert
The Expert module is a crucial component of our Human Practice, serving as a guide for the entire process—abstracted as the mitochondrion in the secretory protein pathway, which provides energy throughout.

Through in-depth communication with Professor Yu Ce from Tianjin University, a participant in China’s 32nd Antarctic expedition, we gained insights into the specific application scenarios of antifreeze proteins and antifreeze chassis, which broadened our directions for improving antifreeze performance.

This part provides future iGEM teams with guidance and references for transforming user needs into actionable project goals, demonstrating how multi-scenario feedback and interdisciplinary collaboration can enhance a project's social relevance and practical potential.

4. Industry
The Industry module functions as the "Golgi apparatus" in our Human Practice model, responsible for processing, integrating, and transforming the project, thus facilitating the transition from technical concepts to commercial applications.

Through collaboration with companies such as SNEFE Research Institute and Bloomage Biotechnology Co., Ltd., we developed a synergistic antifreeze system combining trehalose and antifreeze proteins, forming a dual low-temperature protection mechanism that significantly enhanced the product’s cold resistance and competitiveness.

This part offers future iGEM teams guidance and references for designing pathways toward product development and industrialization, illustrating how partnerships with industry can bridge the gap between research and market-ready applications.

5. Government
The Government module is metaphorically represented as the "cell membrane" in our model—it not only defines the ethical, legal, and socio-impact boundaries of the project, but also serves as the starting point for its commercialization and broader societal release.

Through communication with government heating authorities and a comparative analysis of policies across multiple countries, we recognized the environmental and economic potential of antifreeze proteins. As a result, we plan to incorporate them as an eco-friendly alternative into our project design.

This section provides future iGEM teams with a reference for integrating legal and policy perspectives at an early stage, ensuring that technological development and product implementation align with regulatory requirements.

The five sections above summarize only the most representative achievements of each module; further details can be found on the iHP (integrated Human Practice) page.

Our "Secretory Protein" model integrates the logical structure and workflow of all Human Practice modules, clearly demonstrating how Human Practice provides guidance for wet experiments, enhances the visibility of the project in the fields of synthetic biology and iGEM, and aligns with the principle of “coming from society and returning to society.” This comprehensive and systematic approach offers valuable insights and references for future iGEM teams in conducting effective Human Practice.

[1] Boder ET, Wittrup KD. Yeast surface display for screening combinatorial polypeptide libraries. Nat Biotechnol. 1997;15(6):553-557. doi:10.1038/nbt0697-553
[2] Lin Q, Qi H, et al. Robust orthogonal recombination system for versatile genomic elements rearrangement in yeast Saccharomyces cerevisiae. Sci Rep. 2015;5:15249. doi:10.1038/srep15249
[3] Li J, Gong S, et al. Creation of a eukaryotic multiplexed site-specific inversion system and its application for metabolic engineering. Nat Commun. 2025;16:1918. doi:10.1038/s41467-025-57206-w