Overview

After conducting the previous modeling, we now obtain the possible structure and the binding position of the fusion protein. The next step is to inspect on the fusion protein’s enzyme activity. We want to know the degrading efficiency under some circumstances, hence by constructing a degradation model, and expect to learn how well does the fusion protein work. Finally, we sincerely send our gratitude to OUC-China on assisting us using Gillespie algorithm to process Stochastic simulation.

Result

Reaction formula

ODE equation

Fit the experimental data:

1.  Estimation method: Estimating non-mixed effects model with Lsqnonlin.

2.  Estimated parameters:

•  Raw data:

Fig. I Experimental data of degradation of aflatoxin.

Fig. II Data comparison between experimental data and simulation.

Fig. III Estimated parameter values and their standard errors. Note that the standard errors are extremely higher than the parameter values, demonstrating that though parameter values shown here are best estimated, they are not accurate for lacking adequate experimental data.

Stochastic simulation

1.  Simulation method: Gillespie algorithm

2.  Number of samples: 30000 (300shown)

Fig. IV Stochastic simulation results of degradation of aflatoxin by MSMEG_5998. Red lines represent the stochastic simulations, and the black line represents the mean level of degradation.

3.  Comparison between experiments and simulations

Fig. V Comparison between experimental data and simulation results. Simulation data is the mean level of the stochastic simulations, R^2 is calculated to be 0.9962.

Discussion

1.  According to the raw data, the fusion protein is able to degrade Aflatoxin B1, which matches the prediction from the results we learned in the docking model.

2.  Via the stimulation, the team is able to learn after a certain period of time that how much Aflatoxin B1 will be remaining. According to the results, the fusion protein is able to degrade 50% of the toxin within 3 hours and reach a 70% degradation in 7 hours. It is an amazing information to our team that how great the ability of degradation of syn-MSMEG5998 is.

3.Since the cofactor, F420 is not easily accessible and the control variables are complicated, it is difficult to construct an enzyme model by fully working in wet lab. Fortunately, with the modeling techniques, the team is able to systematically conduct analysis under limited data and resource; therefore modeling enables us to predict results for wet lab and assist the whole project.