viSNE is an excellent method for distilling high parameter cytometry data down to two dimensions to assist with the categorization of events into biological populations. A typical way of categorizing populations within a viSNE map is to use gates:
(areas of a viSNE map categorized into populations using gates)
To learn more about this process, read our article about gating a viSNE map. The problem with applying manual gates to a viSNE map is that it's generally a time-consuming and detail-driven process. Furthermore, it must be repeated for each viSNE run without the ability to transfer existing gates onto new viSNE maps because viSNE results are stochastic.
In thinking about the process of gating a viSNE map, it can be seen how human efforts to gate a viSNE map might be replaced by computer-directed clustering. The clustering approach would find conspicuous regions in the viSNE map and then assign them to groups automatically. There are multiple approaches that can be used for clustering a viSNE map, one of which is running SPADE on the viSNE map itself. The end result of running SPADE on a viSNE map is a collection of clusters that correspond to spatial locations on the viSNE map.
Directions for Running SPADE on viSNE
1) Clone the viSNE analysis
In order to run SPADE on viSNE results, first navigate to a viSNE result. Within the experiment that houses the viSNE result files, click to run a SPADE analysis. Currently Cytobank forces the experiment to be cloned such that the viSNE experiment, which is normally hidden within its parent, becomes a clone that is visible in the inbox.
(click to run a SPADE analysis within a completed viSNE analysis)
2) Create a new SPADE
Within the resulting experiment, do the same operation again to create a new SPADE analysis. This time it will ask for a name and proceed normally to the SPADE setup page.
3) Configure the SPADE run
There are a variety of configurations for a SPADE analysis.
The files being included for this SPADE analysis were previously created by a viSNE analysis. Thus, the ungated population actually corresponds to the population that was previously chosen for viSNE. For that reason, simply choose ungated for the SPADE analysis. A more restrictive population can be chosen if desired for some other workflow objective.
Choose only the two tSNE channels. This application is for clustering the tSNE map only and thus other channels should not be included for the clustering step.
(choose the tSNE channels of the viSNE map to be clustered)
The typical logic applies for choosing fold change groups and baselines. This is a useful way of getting fold change visualizations for a viSNE map, which is usually not possible due to the single cell nature of viSNE results.
Number of Clusters (Nodes)
The number of clusters may need to be honed empirically. A starting point could be to aim low with a number that seems reasonable for the amount of population centers visible in the viSNE map. If there are small/rare populations present, however, the number of clusters may need to be increased to ensure that the small populations wind up in their own cluster as opposed to absorbed into a cluster with other events.
Downsampled Events Target
Downsampling a viSNE map does not necessarily have the same motivational underpinning as downsampling high dimensional data as in a typical SPADE analysis. Avoid excessive downsampling and perhaps avoid it altogether by setting the target to 100 (a target of 100 means 0 downsampling).
Analysis of SPADE on viSNE
Analysis of SPADE on viSNE can proceed exactly the same as analysis of a normal SPADE run, including coloring by channel, bubbling, fold change analysis, statistics, exporting FCS files based on bubbles, etc. The way in which the SPADE tree was created is different, but the analysis follows the same principles.
Colored Cluster Overlay
It can be informative to visualize each clustered segment of the viSNE map by color. This can inform the quality of the SPADE clustering. The colors should correspond to intuitive spatial groupings on the map. Seeing spatially distinct viSNE populations colored the same as a different nearby population means that the populations wound up in the same cluster, and perhaps shouldn't be. Consider using a manual gate for these anomalies and otherwise keeping the other cluster results.
To make this figure, start in the SPADE result and draw a single bubble around the entire SPADE tree. Next, export the bubble as new FCS files. The resulting files will have a Cluster ID channel that will allow the visualization of the viSNE map colored by cluster using colored overlay populations.
Note that the cluster gating workflow that enables this visualization can be easily changed into a heatmap or any other normal analysis method.