aggregateData |
Aggregates results from the clustering and merging step. |
BSPLeaveCenter |
Finds N Leaf centers in the data |
clusterNames |
Get the top features that differentiate the clusters based on the maximum cluster signal levels across all samples. The maximum signal levels of all features of each cluster are found, and then the maximum signal levels are normalized across clusters. Next, within each cluster, the ranks of the features are obtained based on the normalized signal levels. Top ranked features are used to annotate each cluster |
clusterPropHeatmap |
Make heatmap illustrating the cluster proportions across multiple samples. The aggregated data are first padded to assign size of 0 to missing clusters in some samples. Next, the numbers of events in each cluster in each sample are obtained. These values are normalized across samples to find the cluster proportions by samples. The higher the cluster proportion in one sample, the more specific the cluster is to that sample. |
filteredResults_JSON |
Obtain the input to zoomable packed circles plot. The clusters are filtered by size. |
fmeasure |
Compute the F measure between the ground truth and the estimated label |
makeList |
Helper function to obtain a JSON-friendly format of a matrix formatted object |
PAC |
PAC (Partition Assisted Clustering) |
packedCircleInput |
Make packed circles plot illustrating the cluster proportions across multiple samples. |
signalLevelHeatmap |
Make heatmap illustrating the signel levels for clusters. For multiple samples, signal levels differ for the same cluster in different samples. Signal levels can be plotted easily using built-in R functions such as mean, median, and max. |