SeqGeq currently (v1.4.0) offers K-Means clustering within the Clustering platform. K-Means is a machine learning algorithm that places events into one of ‘k’ unbiased clusters, where ‘k’ is an integer set by the researcher.
To run the K-Means clustering, select a population of interest within the workspace, and click on the Clustering button within the Analyze tab of the workspace. In the resulting dialog select the parameter on which you want the clustering to run, and choose your desired k number of clusters to calculate:
K-Means will generate a new parameter which separates clusters by integer values. In order to automatically gate these cluster values into populations, its very useful to take advantage of the AutoGateCategorical plugin, which comes installed by default in SeqGeq. To access and run this plugin, simply select your population of interest, navigate to the Workspace tab of the workspace, and select AutoGateCategorical from the Plugins dropdown list there:
Within the plugin dialog select the K-Means parameter and run the plugin:
This will generate populations corresponding to clusters defined by K-Means:
Try running K-Means clustering on your Quality Cells population, using the principal component parameters calculated previously, and color map those clusters onto your tSNE parameters:
Note: Each color within the color mapping here represents one of the K-Means clusters within the data.
Back to product
