Metabolic Network Model

In this section, we constructed a genome-scale metabolic model (GEM). Using OptKnock calculations, we were able to identify gene knockout strategies that would increase the flux of the target compound. Additionally, we utilized an OD curve model to verify whether the knockout strategies we implemented would impact yeast growth.

Figure 2. The connection between the Metabolic Network Model and experiments.
Read More ->

Protein Structure Model

In this section, we constructed a Structure Model for the enzyme responsible for synthesizing the desired product. By performing homologous recombination mutations on the amino acid residues of the rate-limiting enzyme, we achieved a breakthrough in the rate-limiting step of product synthesis.

Figure 3. The connection between the Protein Structure Model and experiments.

Time Series Prediction Model

To determine the light-dark cycle required for the synthesis of fluorescent proteins, we employed a system of differential equations to describe the synthesis process. Ultimately, we identified the conditions of 1000 minutes of light exposure followed by 25 minutes of darkness as most favorable for continuous synthesis of red fluorescent protein (RFP) and green fluorescent protein (GFP). To further validate the model's soundness, we utilized an ARIMA model to analyze the temporal variations in the synthesis rates of the two fluorescent proteins and provided predictions for their synthesis over the next 10 time units.

Figure 4. The connection between the Time Series Prediction Model and experiments.

Micro-Expression Recognition Model

To obtain real-time feedback from users on our perfume products, we have developed a micro-expression recognition model using machine vision technology based on matching algorithms. Through initial model pre-training, we are able to recognize user expressions and determine their preference for the product based on these expressions. This allows us to adjust our marketing and production strategies accordingly.

Figure 5. The connection between the Micro-Expression Recognition Model and experiments.

Measuring Box Model

In order to quickly measure the concentrations of sclareol and santalol we prepared, we used optoelectronic combination to form a measuring box. We only need to use a colorimetric dish to absorb a portion of the sample and put it into it. Based on the standard spectra measured earlier, the concentration of the fragrance can be obtained, which is more convenient than traditional measurement methods. The equipment can be freely disassembled or added to other hardware facilities, with the hope of creating convenient measurement methods for future teams, and has long-term significance.

Figure 6. The connection between the Measuring Box Model and experiments.