Table 2. Comparison of the clustering properties when varying the distance metric, the distance threshold or the clustering strategy.

Analyses were run on 63,863 RefSeq Bacteria using two different distance metrics, either based on the Jaccard Index (JI) or the Identical Genome Fraction (IGF), six different distance thresholds (from 0.8 to 0.9 and from 0.6 to 0.7, respectively), and two different clustering strategies, either direct (JI-D and IGF-D) or indirect (JI-i and IGF-i; see text for details). All pack sizes were 200 and the clustering mode was set to “loose”. RI, Redundancy Index (# groups / # phyla).

Jaccard Index (JI)
Direct strategy (JI-d)							Indirect strategy (JI-i)
threshold	0.80	0.82	0.84	0.86	0.88	0.90	threshold	0.80	0.82	0.84	0.86	0.88	0.90
RI	59	47	35	25	14	10	RI	54	46	35	20	12	4
# phyla	34	34	29	24	19	11	# phyla	34	31	25	22	13	11
# groups	2005	1589	1025	598	268	109	# groups	1845	1430	870	446	151	49
—pure groups	1992	1576	1009	587	261	106	– pure groups	1835	1416	853	434	149	45
– singletons	1201	904	557	325	143	56	– singletons	1727	818	488	242	88	24
— mixed groups	13	13	16	11	7	3	– mixed groups	10	14	17	12	2	4
– paraphyletic	0	0	0	0	0	0	– paraphyletic	0	1	0	0	0	0
– super-phyla	10	10	12	5	2	0	– super-phyla	10	13	9	7	0	1
– polyphyletic	3	3	4	6	5	3	– polyphyletic	0	0	8	5	2	3
Identical Genome Fraction (IGF)
Direct strategy (IGF-d)							Indirect strategy (IGF-i)
threshold	0.60	0.62	0.64	0.66	0.68	0.70	threshold	0.60	0.62	0.64	0.66	0.68	0.70
RI	74	65	58	45	33	31	RI	50	55	44	30	23	11
# phyla	24	24	22	22	22	15	# phyla	31	25	24	24	19	16
# groups	1776	1548	1271	988	719	464	# groups	1536	1369	1061	715	440	176
—pure groups	1758	1530	1271	971	706	456	– pure groups	1514	1345	1042	701	426	167
– singletons	1094	939	755	587	419	260	– singletons	905	784	595	404	219	77
—mixed groups	18	18	19	17	13	8	– mixed groups	22	24	19	14	14	9
– paraphyletic	4	2	2	2	1	1	– paraphyletic	2	3	1	0	2	0
– super-phyla	11	11	13	10	4	1	– super-phyla	17	17	14	10	8	4
– polyphyletic	3	5	4	5	8	6	– polyphyletic	3	4	4	4	4	5