The Cell Type widget contains the set of Parameter Profile plots that represent one Cell Type in the model. In addition, a Frequency Plot appears at the bottom of the widget to provide information about how well the cell type is being modeled.
You can expand and collapse parameter profiles within the widget for better visualization of the data. To collapse a profile that is showing, click the “shrink” icon next to the parameter name on the left edge. To expand a profile that is hidden, click the “expand” icon next to the parameter name.
|
Click to Shrink |
|
|
Click to Expand |
|
When you click on a parameter profile within the collection of profiles, the profile becomes the active profile and its properties appear in the Parameter Profile Properties panel. See Parameter Profile and Parameter Profile Properties panel for more about parameter profiles.
After you click on a parameter profile in the Cell Type Widget, you can use the PgUp and PgDn keys on the keyboard to move to the next or previous parameter profile. This method automatically opens the next (or previous) profile and activates it.
The Frequency Plot provides valuable information for understanding how well the model for this cell type is working. The Y-axis shows frequency of events, and the X-axis shows the Progression axis (normally 100 states). A bold red horizontal line shows the optimal frequency across the progression axis. The optimal frequency is computed by dividing the number of events by the number of states on the Progression axis. In other words, the ideal model would put the same number of events into each state of the progression.
The jagged black line shows the actual number of events that have been assigned to each state for this cell type. For a model that matches the data well, the black line will follow the red line, bouncing slightly over and under the red line. In places where the black line is significantly above the red line, there are too many events being assigned to those states. Conversely, where the black line falls well below the red line, there are not enough events assigned. This relationship helps identify which Parameter Profiles need to be adjusted. In general, when there are too many or too few events in a portion of the Frequency Plot, look for Parameter Profiles with control points and transitions in those states. Examine those profiles to be sure the control points are properly positioned and the other characteristics of the Parameter Profile are appropriate for the model.
The Frequency Plot also shows the Reduced Chi Square (RCS) of the model and the zones that have been defined.
