Step 6

From ProcessDB
Jump to: navigation, search

Test Your Hypothesis Against Your Data

After completing Step 5, Creating a Model of Experiments, or MOE, you are ready for your final step of comparing your experimental data to the quantitative model you have built, which is a mathematical representation of your scientific hypothesis. This step entails solving your model equations for each of your experimental protocols and searching for parameters that can account for your experimental data. ProcessDB has its own Simulator, based on the very fast CVODE, that can be accessed directly from the MOE context menu, which appears when you right click the MOE. If you prefer you can export your model to the Berkeley-Madonna solver by choosing Export to Solver from that same MOE context menu.. In order to work with Berkeley-Madonna you will have to install it on your computer. ProcessDB also can export your model in SBML however not all features of ProcessDB are currently supported by SBML nor are all SBML features supported by ProcessDB.

When you click Solve in the MOE context menu, you are presented with a dialog box asking you to select a parameter set. If this is the first time you've solved this MOE, there will only be one parameter set (the default) to select.

Select a parameter set and click OK.

A new window will open. Its title bar will say Parameter Explorer and will specify the ID of the MOE you are solving.

The equations will solve immediately using the selected parameter set.

If this is your first solution, a single default graph plotting a few of your model variables will be displayed. You can create as many graphs as you wish and all of these graphs will be retained for the next time you solve this Model of Experiments.

The Parameter Explorer window gives you access to and control over three groups of capabilities. These groups are assembled in the menu bar. From left to right they are:

Graphs

Optimizer

Parameter Sets

On the Graphs menu you can Add a new graph to your collection, Remove a selected graph permanently, choose (using the Ref Plots toggle) whether or not to display solutions for a parameter set you have designated as the reference parameter set, choose whether to display All Graphs or just a subset of graphs you have selected, and choose to Import the graphs defined by a collaborator.

On the Optimizer menu you can open the Edit Associations and Weights screen where you can make the connections between model and data, or you can open the Optimize screen where you can set up global parameter optimization using ProcessDB's particle swarm optimizer or gradient optimization using the orthogonal distance regression method as implemented in ODRPACK95.

On the Parameter Sets menu you can create a New parameter set, Open another existing named parameter set selected from a list and see its solution, Copy the current active parameter set and name the copy so that the parameter set you copied from becomes a saved point you can always return to, use Open Ref to load an existing named parameter set as a reference whose solutions can be displayed on the same graphs as the solution for the current active parameter set, use Make Ref to make the current active parameter set the reference parameter set, and use the All Parameters toggle to display all model parameters or only a selected subset of model parameters that are your current focus.