Why This Project——Solve Their Problems
Before we started the brainstorm for our project, we carried out a survey of iGEM teams about the difficulties they meet in their project. The majority of these teams told us that it's a big challenge for them to select the most suitable ones from large amounts of reactions to construct the metabolic pathway for their products. In the beginning, they need to consult lots of papers and move across several metabolic websites to seek and collect related information. It costs them too much time and energy. Even if they find out a few different pathways that can meet their needs, it's another problem that which one can work better and keep a high production without pre-experiment.
To solve these problems, our team wants to establish an integrated platform to assist synthetic biologists in completing pathway design more conveniently.
With further research on biosynthesis, computational tools will gradually replace some simple but tedious work, and then our software will show more practical potential.
How We Start——Inspiration Inside IGEM
In order to determine the specific function of our software, we investigated many representative iGEM projects, and part of their work inspired us:
① Tongji-Software 2019: Their talent framework for pathway search has a high referential value for our project. ② USTC-Software 2018: The excellent performance of their model gives us the confidence to simulate the engineered organism's metabolic system.
figure1. Logos of the inspiration teams
1. build an integrated database with abundant metabolic data in which users can get nearly all the necessary information.
2. construct a one-stop platform for pathway design and recommend the best enzymes in every step.
3. develop a method to perform metabolic simulation so that users can observe the content change of materials in the modified metabolic network.
What We Have Done
For the first goal, we take the KEGG database as the main body and integrate the other four databases into it. After data cleaning and missing data filling, we organize these data into three tables (Enzyme, Compound and Reaction).
As for the pathway search, we adopt a heuristic algorithm——A* base on the project of Tongji Software 2019 to speed up the procedure with the same accuracy. We also remove the polymers which are not available for the calculation to reduce the search space and avoid unfeasible reactions. In addition, we optimize the weights of each index for rating pathways according to the data distribution. Besides, We add a new practical function ——reverse search to help find out all the compounds which can be transformed into the given target compound in the fixed number of steps.
Various methods of metabolic simulation have been proposed before, but most of them are not practical for the application in synthetic biology. To build a more robust, less data-dependent, and more detailed Metabolism Simulation tool, our team put forward a new hybrid metabolism simulation model. This model takes four different simulation methods to process the reactions, which are documented differently.
Above are our work, and each part makes an immense contribution to our software construction.