Overview

Our iGEM project aims to enhance the quality of life and experience for visually impaired women during their menstrual periods. This is achieved by utilizing Saccharomyces cerevisiae to produce various aromatic terpenes, such as β-caryophyllene, linalool, and nerolidol. These compounds, after being produced by yeast, are encapsulated in a material that is then placed inside sanitary pads. When in contact with menstrual blood, the encapsulated compounds are released, allowing visually impaired women to detect the onset of their period through the scent. Additionally, We tend to select target fragrance compounds that also have antibacterial properties.

To support this, we developed three models to simulate key aspects of the process. Firstly, we constructed a yeast growth model based on the data collected from 24-hour growth measurements, fitting the data using a logistics model. Second, we created an aroma diffusion model to simulate how long it would take for the released scent to reach the user. This diffusion model, based on Fick's second law with a dissipation term. Finally, we used AlphaFold to predict the protein structures of the genes involved in producing the desired compounds. Each of these models contributes to refining and optimizing the overall design of the system.

The Yeast Growth Curve Model

The atmosphere in the lab was full of anticipation, with small vials of yeast cultures spread across the table. We had just constructed plasmids containing genes for the synthesis of different aromatic compounds and successfully introduced them into Saccharomyces cerevisiae. To track the growth of these genetically engineered yeast, we decided to collect samples every two hours over the next 48 hours to record the yeast growth curve.

We saw the OD600 values, indicating changes in yeast cell density, and this data formed the basis for our subsequent modeling.The resulting data is presented in the picture below.

We also measured control group data for comparison, and each sample with three biological replicates.

Finally, we modeled the growth curves using the logistics model based on the collected data.

Lin Chuqiao : "Come take a look, the OD600 values for these samples are rising rapidly. The yeast growth seems very active."

Tian Jiameng : "These data look very clear. I can fit a logistic model with them. This model will help us better predict yeast growth dynamics under different culture conditions."

Through data analysis, we found that the yeast entered the rapid growth phase between 6 and 20 hours, with OD600 values rising sharply, indicating a significant increase in cell density. Then, between 20 and 30 hours, most samples slowed their growth, with OD600 values plateauing as the yeast entered the stationary phase.

Chen Quqing :"Did you notice? The groups with the POT plasmid have almost identical growth curves, with no significant differences."

Lin Chuqiao : "It seems that the POT plasmid has a similar effect on these groups, but there are significant differences in growth under different culture media. We can use this data to optimize the conditions for the next experiments."

We also conducted comparative experiments to evaluate the effects of different culture media and whether the shuttle plasmid was introduced. The results are shown in the Figure 3. Our results indicated that yeast strains carrying shuttle plasmids exhibited comparable growth in drop-out medium medium, with an OD600 of approximately 4.5. In contrast, the yeast chassis strain reached an OD600 of about 8.5 in YPD (Yeast Extract Peptone Dextrose Medium) medium. This discrepancy suggests that the composition of the medium significantly influences strain growth.

The model allowed us to assess the efficiency of yeast growth in different culture media, identifying optimal conditions for rapid proliferation.Based on the growth curves, we can optimize the conditions for further experiments. The insights gained from this model will serve as a foundation for refining our experimental design, helping us better manage the trade-offs between yeast growth and the biosynthesis of target compounds like terpenes.

Aroma Diffusion Model

Several team members were gathered around a table, discussing the problem at hand.

We were working hard to solve a key challenge--how to ensure that aromatic terpenes could diffuse into the air in time for visually impaired women to detect them.

Tian Jiameng :"We need to build an aroma diffusion model to simulate the release and diffusion of aromatic terpenes from the sanitary pad material into the air. This is critical for determining how long it takes for the scent to be detected by the user."

We decided to use Fick's Second Law to simulate the diffusion behavior of these aroma compounds, such as β-caryophyllene, linalool, and nerolidol. These compounds are encapsulated in a material that releases them upon contact with menstrual blood. In this model, we considered the diffusion coefficient of the aroma compounds in the air, as well as their natural dissipation over time.

Chen Quqing:"I've already integrated Fick's Second Law with a dissipation term and simulated the aroma diffusion using Python.

The complete model equation is shown below:

In this model, several important parameters are derived from the following sources:

Diffusion Coefficient (D):We referenced the typical air diffusion coefficient for terpenes such as limonene, which is around 1e-5 m²/s. Therefore, we set the diffusion coefficient DDD to 1e-5 m²/s

Dissipation Rate (r):The dissipation rate is set to 0.001, representing the proportion by which the aroma compound's concentration decreases per second.

Distance (L):The distance from the release point of the aroma compound to the user's nose is set to 0.5 meters.

Olfactory Threshold：The olfactory threshold is the minimum concentration at which a person can detect an aroma compound. We set it to 0.016 mol/m³, based on the olfactory threshold for limonene.

As the simulation results came in, we could clearly see the diffusion and dissipation process of the terpenes. According to the simulation, the scent becomes detectable after about 15 minutes, peaks at 30 minutes, and fades away after 55 minutes. This model not only helped us predict the release and diffusion speed of the aroma but also allowed us to optimize the release mechanism to ensure users detect the scent at the right time.

Tian Jiameng: "This is exactly the result we needed! The model provides solid data support for designing the most effective aroma release system."

With the precise mathematical simulation provided by this model, we could evaluate all aspects of the aroma diffusion process, from the release of aroma compounds to their detection by the user. This allowed us to more efficiently adjust the aroma release rate and also provided the possibility to reduce experimental costs—we could test and optimize different parameters through the model before conducting real-world experiments. Ultimately, this will help us design more efficient aroma release mechanisms.

Lin Chuqiao: "By running simulations, we've saved significant time and resources, and we can test and optimize key parameters before actual experiments, which dramatically reduces costs."

In the future, we will conduct experiments to validate the data obtained from the model and assess the product's actual effectiveness. By integrating experimental data, we can refine and optimize the current aroma diffusion model to ensure its accuracy and reliability in real-world applications. Through testing the diffusion speed under different conditions and user response times, we will gain a deeper understanding of the product's performance in practical environments, allowing us to further enhance product design and user experience.

Protein Structure Modeling

In the laboratory, our team members sat around the computer, focused intently on the screen.

On the display were the sequences of 9 key genes we had just uploaded. Each of these genes encoded proteins crucial for the synthesis of the target terpenes in our project. We had decided to use AlphaFold3, a powerful tool capable of rapidly and efficiently predicting the 3D structures of proteins. With a click of the mouse, the gene sequences were entered into the AlphaFold server, and all we could do now was wait.

A few hours later, the predicted 3D protein structures appeared on the screen, and the team couldn’t help but let out murmurs of excitement.

Lin Chuqiao: "Look at this active site! Its shape is exactly what we expected. This will really help us design more effective enzymes!"

Chen Quqing: "Yes, and the high confidence scores of these structural models give us even more confidence in our experimental design. We can start thinking about how to use directed evolution to optimize these proteins."

These complex 3D structures not only revealed the key functional domains of each protein but also gave us deeper insights into their roles in terpene synthesis. These predicted structures provided a solid foundation for future optimization strategies, particularly in enhancing the functionality of active sites and improving the overall performance of the proteins.

Conclusion: Integration of Models and Future Outlook

Through the development of these three models—protein structure modeling, yeast growth curve modeling, and aroma diffusion modeling—we have systematically addressed the core challenges of our project. These models have provided valuable guidance and optimization strategies, from enhancing yeast metabolite production to ensuring the efficient release of aromatic terpenes, with each step backed by precise simulations and data analysis.

The yeast growth curve model enabled us to predict yeast growth dynamics under different conditions, optimizing the production of terpenes.

The aroma diffusion model ensured that we could design the best aroma release system for users, providing an optimal user experience.

Protein structure modeling allowed us to deeply understand the functional domains of our target proteins, laying the groundwork for optimizing enzyme efficiency through directed evolution.

The integration of these models not only accelerated our experimental progress but also reduced costs, allowing us to achieve significant results in a short time. Moving forward, we plan to validate these models with real-world experiments and further refine our designs. Our future work will not only involve improving the models but also turning these theoretical results into practical products that enhance the quality of life for visually impaired women.

We firmly believe that through continuous effort and innovation, we will ultimately be able to translate our research from the lab into real-world applications, helping a wider population.

Reference

[1]Cakir A,Kordali S,Kotan R.Screening of antibacterial activities of twenty-one oxygenated monoterpenes[J].Zeitschrift für Naturforschung, C. A Journal of Biosciences.2007,62(7)

[2]Zeng W, Jiang Y, Shan X, Zhou J. Engineering Saccharomyces cerevisiae for synthesis of β-myrcene and (E)-β-ocimene. 3 Biotech. 2023 Dec;13(12):384. doi: 10.1007/s13205-023-03818-2.

[3]Limm W, Begley TH, Lickly T, Hentges SG. Diffusion of limonene in polyethylene. Food Addit Contam. 2006 Jul;23(7):738-46. doi: 10.1080/02652030600654408. PMID: 16751151.