Interaction Plots

The Interaction plot is used to visualize values of a feature for groups considered by a statistical test. An interaction tests for whether the effect of one factor is dependent on another factor. For example, in an experiment where drug-resistant or drug-sensitive cell lines are treated with either vehicle or drug, we would expect the effect of the drug to depend on whether the cell line is drug-resistant or drug-sensitive; to account for this, we would include an interaction between cell line and treatment in the statistical test.

The Interaction plot is available in Feature list data nodes generated by ANOVA, GSA, and Hurdle model differential analysis tasks when two categorical attributes are included as factors in the statistical test. For ANOVA and Hurdle model, an interaction between the two factors must also be included.

Click the View interaction plot button in the View column of the task report

The x-axis of the interaction plot is one of the categorical attributes or factors included in the interaction. The y-axis is the LSMean of the input values for a feature. The points represent the groups formed by combination of the two factors included in the interaction.

The points on the plot are colored, shaped, and connected by the factor in the selected interaction that is not being used as the x-axis.

Configuring the Interaction plot

The interaction plot can be configured using the control panel to the left of the plot

To switch between interactions included in the statistical test, use the Interaction drop-down menu.

To switch the factors being used to group and color/shape the points, use the X-axis drop-down menu to change the x-axis grouping factor.

Click < Previous and Next > to move up and down the table.