Algorithm 1.

SCORE Algorithm

1 Create a dendrogram tree using CORE, keeping Euclidean distance matrix M for bagging runs:

2 Create a vector bk (k = 1, 2, …, m);

3 Populate bk with a subsample of cells, with replacement, from the set of all cells $\mathcal{C}$;

4 Create a new matrix, Nk, of Euclidean distances for the cells in bk, using values from M;

5 Generate a new dendrogram tree and clustering of cells;

6 Record result from optimal stability of subsampled tree;

7 Vote on most commonly occurring result;

8 Choose most stable result from the original dendrogram tree