Table 1 |.

Number of clusters $(k)$ and ARI, comparing outputted cluster labels to the ground-truth population labels, when our clustering pipeline, sc-SHC, as well as existing methods, are applied to three datasets. We also applied our significance analysis on top of each existing method

	One-population				Five-populations				Five-similar-populations
	Original		With significance analysis		Original		With significance analysis		Original		With significance analysis
	$k$	ARI	$k$	ARI	$k$	ARI	$k$	ARI	$k$	ARI	$k$	ARI
sc-SHC	1	1.00	N/A	N/A	5	1.00	N/A	N/A	5	0.99	N/A	N/A
Seurat-Louvain-Clustree	6	0.00	1	1.00	5	0.99	5	0.99	4	0.78	4	0.78
Seurat-Louvain-Default	9	0.00	1	1.00	7	0.89	5	0.99	8	0.67	4	0.78
Seurat-Leiden-Clustree	4	0.00	1	1.00	8	0.88	6	0.93	4	0.78	4	0.78
Seurat-Leiden-Default	9	0.00	1	1.00	8	0.88	6	0.93	10	0.59	4	0.78
Monocle	2	0.00	1	1.00	6	0.96	6	0.96	4	0.74	4	0.74
SC3	15	0.00	1	1.00	11	0.91	8	0.96	11	0.91	7	0.97