Team:NCTU Formosa/Model
- ☰
Overview
Modeling is calculating the physical phenomenon by using mathematical methods or logical algorithms. Our model aims at computing mutation rate from bacterial growth curve and predicting the function of quorum sensing system. Besides, we also develop an AI which can accurately predict the mutagenicity. Our model can be divided into the following three parts:
First, computational growth model simulates and analyzes the result of our experiment. Mainly focusing on the function of substance toxicity, nutrient limit, and in the end, we can get the mutation rate from the growth curve.
On top of that, quorum sensing model predicts the effect of mutagenicity to the time of red fluorescence.
Furthermore, we develop an AI to predict the mutagenicity by analyzing the chemical structure. Take our database of chemical structures and Ames test results as training data for machine learning, and then we use support vector machine (SVM) algorithm to train our model to accurately predict the mutagenicity.
All in all, for unknown chemical structure, we can measure it by our engineered E. coli, and calculate the time of red fluorescence to get our quantified mutagenicity. For known chemical structure, we can quickly get the mutagenicity by using the result of machine learning.
Figure 1: The workflow of our model
Go to >>>
