Table 3. Comparison of results from benchmark jobs.

There are a total of 291 models in the benchmark suite. Protocol descriptions are shown in Table 2 ▶. Any, solved by any protocols; All, solved by all protocols. Solved, a solution with MapCC 0.2 (with respect to the target structure) has been found. Identifiable, a solution with TFZ 7.0 has been found. Gain: average fractional difference in log-likelihood gain with respect to protocol 1.

	LSQMAN (strict)			LSQMAN (tolerant)			ClustalW
Protocol	Solved	Identifiable	Gain	Solved	Identifiable	Gain	Solved	Identifiable	Gain
1	183	126		173	119		159	102
2	182	127	0.042	174	115	0.003	163	104	0.001
3	182	124	0.021	173	112	0.002	165	102	0.005
4	178	127	0.004	179	123	0.007	160	99	0.011
5	182	126	0.018	173	116	0.007	160	101	0.012
6	185	129	0.035	182	127	0.026	164	100	0.007
7	182	129	0.006	181	118	0.016	162	103	0.002
8	182	134	0.033	177	123	0.031	158	108	0.050
9	185	129	0.009	173	122	0.028	165	109	0.056
10	188	135	0.093	180	128	0.096	162	113	0.098
11	186	133	0.066	181	130	0.089	162	113	0.100
12	188	136	0.092	178	124	0.114	163	110	0.113
Any	199	146		190	137		183	123
All	165	111		155	99		134	87