1 Comparison Correlations

Comparison Correlations

Overview

The Dynamic Correlation - Comparisons Correlation view is a fast and easy way to identify genes that either are co-regulated in different conditions (i.e. Disease vs Normal, Responders vs. Non-responders, Treatment vs. Control) or have an negative fold-change correlation in various studies. This view is dynamically updated to reflect users' filters based on project or comparison metadata.

Pearson correlation is the method used for analysing the data.

Usage

After searching a gene, this view will be available under select view:

Filtering

All project and comparison metadata can be used to filter comparisons of interest, such as liver disease in the example below:

Actively used filters will appear red to indicate filtering has been done.

Output

As shown in the picture above, the resulting output will be divided in two windows: (1) The correlation table on the left, which shows the top correlated genes, sorted by value of correlation (positive correlation and negative correlation of all genes in the comparisons) and (2) a view that plots the comparison value of the queried gene on the x-axis (in this example, SERPINH1) and correlated gene on the Y-axis (in this example, THBS2).

Correlation Table

Clicking within the window with the correlation results will instantly update the view on the adjacent panel. In the Correlation Table window, users will see columns for the Gene, correlation statistic, and GeneID). For more details of the correlation, users can click the "View Correlation Table" at the bottom of this view.

This table will provide a full description of the genes in the table, along with the correlation value.

Correlation View

This view allows users to visualize the correlations found in the Correlation table. The input data (queried gene) will be plotted on the X-axis and the correlations plotted on the y-axis. As in all other views in land, this view can be customized using the Task tab of the view controller on the right side of the screen.