Heat Map

Heat map overview

CosmosID provides a heat map for visualizing comparative analysis results. Heat maps are visualizations that use color to represent values and allow us to see relationships in the data. Cells in our comparative analysis matrix are colored based on a red to blue spectrum with red representing high relative values and blue representing low relative values. For relative abundance as the value used, for example, we would see red cells for organisms with high abundance, yellow cells for those with medium abundance, and blue cells for those with low abundance.

Why use a heat map?

Heat maps are great for looking for trends in large datasets, whether a large number of samples or a large number of organisms among the sample (or both). For example, if you have two cohorts of samples and see some organisms are consistently highly abundant in one cohort but low in another cohort this might help you identify certain organisms or taxa for further study.

How it works

The heat map shows the relative intensity of a value within the matrix. The numbers that are highest are given a “hot” value - more orange or red and the numbers that are lowest are given a “cold” value or more blue. The colored cells are plotted on a graph, allowing you to identify samples or taxa of interest.

Options for viewing the heat map

Scale - the scale at the top shows the color ranges for the given values chosen for your comparative analysis.

Most Abundant Taxa - you can adjust field to include the top percentage of results. For example, if I enter 95 into the "Most Abundant Taxa" field, the bottom 5% of calls will not be shown in the heat map.

Clustering - You can adjust the clustering of your samples to improve your visualization and interpretation. This feature can allow you to realize, discover or validate important relationships.

Sample Clustering is the default. Samples are placed on horizontal axis and clustered based on compositional similarity. Taxa are shown on the vertical axis and ordered high to low in regard to mean abundance.

Cohort Clustering allows you align your samples by cohorts defined when creating your comparative analysis alphanumerically. Taxa are shown on the vertical axis and ordered high to low in regard to mean abundance.

Taxon Clustering allows you to align taxa on the vertical axis by similarity. Cohorts are ordered alphanumerically.

Mouse hover - if you hover your mouse over the heat map it will show the organism called in that cell, the actual value represented in the heat map, and the label that the sample belongs to, if applicable.

Scroll- if the heat map is larger than the screen, use your mouse to scroll up and down while hovering over the heat map.

Zoom- The user can left-click and drag creating a highlighted area. The visualization will zoom into this area to allow closer investigation. Double-Clicking will reset the visualization prior to zooming in.

Fit-To-Screen- The fit to screen option alters the visualization to fit the user's screen. This is useful when processing a large number of metagenomic samples. The visualization by default does not always fit within the page and this allows a concise view.

Export - to download the heat map, click "Export" in the top right corner and select PNG or SVG.

Change taxonomy: To change taxonomic levels for organism databases, you can click on the levels on the top bar:

Logscale: Sometimes it is difficult to discriminate low abundance calls if you have some results that are at a very low abundance and some that are high abundance. In order to better distinguish between low abundance calls, we provide an option to view the log scale version of your data. To view the log base 2 version of your heat map you can do so by selecting "Yes" next to Logscale in the upper menu bar.