|
|
Cluster Heat Map
You can click and drag on gene or sample names to change the displayed region of the heat map. When you move the cursor over a gene or a sample name, you can use the mouse wheel to zoom in or zoom out. You can click a specific gene name to display additional information for that amplicon.
All samples and amplicons are jointly clustered to produce a single clustering result.
Before the clustering process, these normalizations steps are performed:
| 1 | Application of a minimum count threshold of 1. |
| 2 | Log transformation of counts. |
| 3 | Median normalization of counts across samples. |
Agglomerative hierarchical clustering is then performed on the normalized values using correlation as the distance measure along with average linkage. When the clustering process completes, the data are MAD-normalized within each amplicon for display.