Running statistical analysis

Before running PathVisio statistical analysis, you will need the following:

  • Select the correct identifier mapping database.
  • Import your data to PathVisio.
  • Download the appropriate pathway collection that you want to run the analysis on.

Once you have these prerequisites in place you are ready to run statistical analysis.

  1. Select Data -> Statistics to open the Pathway statistics dialog.
  2. Use the Expression field to enter a criterion for running the analysis. This is similar to entering a criterion for a color set. For example, you could specify a criteria based on a fold change threshold. You can either type the expression in the Expression field or you can use the available parameters and operators listed by clicking on them.
  3. Select the correct pathway archive under Pathway directory.
  4. Click Calculate to start the analysis.

Interpreting the results table

When the analysis is finished, the results will be displayed in the statistics interface. The results are formatted as a list of all pathways in the designated pathway archive, with a set of metrics reported for each pathway, explained below. Clicking on any pathway in the list opens the pathway in PathVisio. It is also possible to save the results table as a tab-delimited text file by clicking the Save results button.

Metrics in the results table:

  • positive (r): The number of genes on the pathway that pass the criterion.
  • measured (n): The number of genes on the pathway that were measured in the experiment.
  • total: The total number of genes on the pathway.
  • %: The percentage of genes measured that passed the criterion (positive/measured).
  • Z Score: The standard statistical test under the hypergeometric distribution. A high Z score indicates that the pathway has more genes that pass the criterion than would be expected by chance.
  • permuted p-value: A permutation test is performed to calculate permuted p-values. The data is permuted 1000 times and a rank is calculated of the actual Z score compared to the permuted Z Scores. Pathways with a permuted p-value < 0.05 are significant.