Table 1.

Experimental results comparing the performance of the SPR supertree method to RF and MRP supertree methods

Data Set	Supertree Method	SPR Distance	RF-Distance	Parsimony Score	Time (s)
Marsupial (267 taxa; 158 trees)	SPR	382	1604	2203	1097.79
	SPR–RF–TIES	373	1536	2149	767.01
	SPR–MRP	380	1534	2126	219.64
	RF-Ratchet	386	1510	2142	688.55
	RF–MRP	381	1502	2118	662.95
	MRP–TBR	379	1514	2112	20.52
Seabirds (121 taxa; 7 trees)	SPR	17	109	235	31.15
	SPR–RF–TIES	17	63	208	29.44
	SPR–MRP	17	61	208	2.04
	RF-Ratchet	17	61	210	6.34
	RF–MRP	17	61	209	5.87
	MRP–TBR	17	61	208	1.03
Placental mammals (116 taxa; 726 trees)	SPR	1715	5908	8946	5561.84
	SPR–RF–TIES	1713	5902	8934	5040.03
	SPR–MRP	1713	5876	8921	1819.08
	RF-Ratchet	1784	5718	8830	442.697
	RF–MRP	1781	5694	8820	430.77
	MRP–TBR	1783	5702	8809	34.27
Legumes (558 taxa; 19 trees)	SPR	108	651	1175	21130.08
	SPR–RF–TIES	92	471	1037	12376.00
	SPR–MRP	110	511	903	276.49
	RF-Ratchet	126	409	1095	403.513
	RF–MRP	136	451	1081	397.62
	MRP–TBR	140	519	891	579.76

Notes: Six analyses are shown: The SPR supertree method starting from an SPR greedy addition tree (SPR) or MRP supertree (SPR–MRP), the SPR supertree method breaking ties with the RF distance using a greedy addition tree (SPR–RF–TIES), the RF supertree method starting from random addition sequence trees (RF-Ratchet) or MRP supertree (RF–MRP), and MRP with TBR global rearrangements (MRP–TBR). The best optimization criteria or running times for a data set are shown in bold.