In a correlation analysis, you are using statistical correlation to assess the relationship between variables.
Data input requirements:
- Only one cohort is supported.
- You have to add two numerical nodes.
- Categorical nodes are optional to create coloured groups.
After fetching data:
First the method for computing the correlation and a data transformation setting have to be selected.
Options are: Pearson, Kendall, or Spearman, and raw, log2, or log10 respectively.
The default view after creating the plot shows a scatter plot with the two selected nodes. Every dot represents an individual, with details shown on hover over. On the axes bins are shown with counts for that specific range. A line is drawn that represents the calculated correlation and intersection. Details are shown when hovering over the line. On the right some basic statistics are shown.
Using the mouse, you can select a subgroup of individuals to recompute the basic statistics on the right. Also the correlation will be recomputed and redrawn. The selection box you’ve created can be dragged. Right clicking it gives the option to zoom in on that area, to remove those individuals from the computed statistics, or to reset the entire selection.
You display values as coloured dots instead of black by including categorical values in the Fetch data step.