Team:Jilin China/Model

MODEL


Background Methodology Integration

Model

  • Background

    Since RNA-based thermoregulation performs in a more immediate and more effective manner, we focused on building a RNA-based thermosensors toolkit. The principle of heat-inducible RNA-based thermosensors is when temperature increase, hydrogen bond fracture, and the secondary structure (stem-loop structure) of RNA thermosensors is destroyed. The RBS sequence exposes and fluorescence increases. However, lacking of a proven Toolkit hamper the heavy use of this kind of thermosensor. Our model is aiming at building a data base for our Toolkit with focused features for potential users.

    Figure 1. principle of heat-inducible RNA-based thermosensors

  • Achievement

    • ·Developed a Curve Fitting Model to extract focused feature of Themosensors.
    • ·Developed a quantitative criterion to judge an applicable thermosensor.
    • ·Developed an Intelligent Screening System to avoid undesirable thermosensors, improving our experimental efficiency.
    • ·Found free energy and GC content as crucial factor on RNA-based thermosensors.
    • ·Developed an Thermo Loss System to evaluate the cost effect of temperature.
    • ·Made our script publically available for use by other teams.

  • Modeling

    1.The Curve Fitting Model bridges the gap between experimental data and focused thermal features for potential users. This model provided us a continuous fitting curve to describe the switch behavior of our RNA-based thermosensors in detail. Statistical Postulate on two-state distribution, Thermodynamics, Logistic Regression and least square method were used to fit the curve. We then extracted features from the curve to characterize each thermosensors. Besides, the model provided us quantitative criterion R-squared to judge an applicable thermosensor, which was employed to generate training sets in our Screening System. (see more)

    2.The Screening System is an intelligent machine learning system to screen the desirable RNA-based thermosensor. Using the random forest algorithm to minimize undesirable samples, we predicted the validity of thermosensors by the sequence and get feedback to improve our Toolkit construction efficiency. (see more)

    3.The Thermo Loss System built a system to estimate the heat dissipation of Yogurt fermenters. Using heat conduction equation, Newtons heat-transfer law and classical difference method, we simulate the heat dissipation and analysis the cost saving when 1℃ is decreased on fermentation temperature. (see more)

    Experiment Switching Behavior Database of RT Toolkit Thermal Intelligent Screening System Data Focused feature of RNA Thermosensor Validity Index of RNA Thermosensor Feedback:Design of Experiment to get Screen RNA Sequence with High Validity Melting Temperature Human Practices Ask for melting temperature, relative intensity, sensitivity.

  • Integration

    Our models are closely interlinked with our project. Results of our model also feedback to our design and human practices work in several important ways. Some examples include:

    1.The Curve Fitting Model bridge the gap between experimental data and valid thermal features for potential users. It provided us a detailed and continuous description of behavior of our RNA thermosensor by fitting a temperature-dependent expression curve. Moreover, it provided us a quantitative criteria to judge a valid thermosensor.

    2.The Screening System is an intelligent machine learning system to screen the valid RNA sequence. Using the random forest algorithm, it feedback to the design by minimizing invalid samples.

    3.The Cost Analyses built a heat conduction equation to estimate the heat loss power of Yogurt fermenters, providing us an important commercial reference to learn the effect of our production on cost saving and energy conservation during the work of human practice.