In our design, we aim to manufacture psicose, which has a lot of advantages over other sugar or sweeteners as is mentioned in the research of the psicose. The advantages of psicose include low energy, benefit to diabetes and hyperlipidemia, making psicose become more and more popular in people. People can get a psicose sweetener or food rich in psicose, in which psicose is synthesized by biological ways, because using chemical ways is not a good choice to synthesize clear and edible psicose in a food grade for the containing of unhealthy and poisonous by-product in chemical synthesis methods.

Nevertheless, biological methods in producing psicose are inefficient due to the low enzyme efficiency of D-psicose 3-epimerase. The efficiency of D-psicose 3-epimerase is different due to the various environment system and the different concentration of substrate (sounds amazing due to the efficiency of enzyme irrelevant in most occasions).

Our aim is to model our project and predict the situation before wetLab or in the future. Also, it is necessary to validate our model and prove that our model is available.

The first model we established is the dynamic models for the four devices. The concentration of reagent and product can be measured by wetLab and the results are the same with wetLab and the model from SYSU-Software, which validate our dynamic model by sharing our wetLab data.

The second model we made is the production simulation model and conversion rate model. We use probability theory to simulate gene mutation process and the directed evolution. Thus, we get the production prediction and the conversion rate prediction, as well as the time we need to improve an enzyme.

The third model is the hairpin coupling model. We cooperate with OUC-China and set up hairpin coupling model with the theory of statistical thermodynamics and gene expression. Thanks to the help of OUC-China, we can set this model and predict the hairpin coupling coefficient.

The forth model is market prediction model. We set up this model based on HP data from the survey, which makes our model more close to the truth.

Our four models have good portability, making it possible to use in other system to deal with production simulation, hairpin coupling and market prediction. On the other hand, the basic data is needed to initialize the models and get some necessary parameter, which is OK either from paper or from wetLab.


Microfluidics deals with the behaviour, precise control and manipulation of fluids that are geometrically constrained to a small, typically sub-millimeter, scale at which capillary penetration governs mass transport. This year, we designed a microfluidics-based biological application named “Christmas Tree” to be a platform of reaction to achieve concentration gradients automatically and save reacting time. Besides, we optimized the device to be detachable by by adding conical channels at the bottom of the reaction well, so that we can do parallel experiment by changing rows of well-plates with different OD values of bacteria. In addition,we can add pressure like voltage to control the velocity of the fluid and design various sizes of channel to achieve the concentration gradients we need through fluid simulation with application, ANSYS.