This function provides a different way to view and explore data. Apart from viewing many experiment tracks over same genomic regions, you might want to focus on one experiment and view the behavior of your list of favorite genes in this experiment. Gene Plot helps you do it.
After you run the Gene Set View, you can see a orange button in the control panel:
Hit that button, the Gene Plot panel will be displayed on top of it:
After a custom bigBed track has been submitted, right click on the track image, in the context menu you will find the "Gene Plot" option:
Selecting this option will launch the same Gene Plot panel as above. Then you can run Gene Plot using items in the custom bigBed track. However this function is still experimental and please let us know when it runs into error.
You need to go through 3 configuration steps to get your ideal Gene Plot:
1. select heatmap track The heatmap track provides data to be plotted. In the drop-down menu lists all tracks on shown in genome heatmap, both native and custom. You can make convenient selection via the right-click menu option.
2. choose graph type Click a tab in the center box to choose that plot type. All plot types will be discussed at later sections.
3. choose rendering method Usually it is recommended to use Google Chart Tools, which will generate nice interactive diagrams. In case that's not applicable (too much data, or service unreachable), you can select R software on the Browser server to render a still PNG image.
After running Gene Plot, a new panel containing the plot will be displayed above this current one. Click ✘ to dismiss.
In its panel you can find brief explanation and a few controls. Most important one is "Number of data points". An example plot is following with 50 for data point #:
Change it to 200, the curve will show finer detail:
Similar as first plot type, but now each gene/item is plotted in one curve:
In this type genes are dissected into 5 types of components, and each type is plotted as an average curve over all genes. This plot style only applies to genes:
This is a lightweight implementation of hierarchical and K-means clustering on the gene sets, and their visualization in heatmaps. The rendering only depends on R software but not Google services. Following show the heatmap and dendrogram from hierarchical clustering analysis:
The appearance of heatmap will look a bit different if you uncheck the "use global max" option. In such case, max value of each item will be used to determine heatmap cell color:
In the case of hierarchical clustering, you can click on juction point of dendrogram. The sub-cluster will be highlighted and genes will be printed out:
During K-means clustering you can define the number of clusters you want to find, and in the result you can get genes in each cluster:
After you run the Gene Plot, in the graph panel you can find two buttons to let you get the data underlining the plot. "plain text" will yield a simple html table under the graph. Following is how it looks like for plot type #3, rows are each gene part type and columns are data point:
The fancier choice is Google Chart table, which look like following with same data.
Last modifed: 1/10/2012