Table 1.

Probability of recovering the correct tree in 1000 replicate datasets with 5000 codons simulated under the homogeneous (homo) model

	Homo, short trees						Homo, deep trees
	Tree 1		Tree 2		LBA Tree		Tree 1		Tree 2		LBA Tree
Sim model	M0	M3	M0	M3	M0	M3	M0	M3	M0	M3	M0	M3
AA-WAG+G	0.84	0.70	0.53	0.44	0.88	0.82	0.32	0.52	0.09	0.32	0.87	0.88
AA-GTR+G	0.80	0.62	0.78	0.63	0.97	0.93	0.32	0.41	0.36	0.55	0.92	0.90
DNA-123	0.81	0.73	0.82	0.70	0.98	0.96	0.45	0.41	0.26	0.28	0.58	0.27
DNA-123-P	0.78	0.67	0.85	0.75	1.00	0.99	0.32	0.30	0.40	0.41	0.97	0.88
DNA-12	0.71	0.48	0.76	0.60	0.98	0.95	0.31	0.28	0.30	0.36	0.84	0.72
DNA-12-P	0.65	0.47	0.78	0.59	0.98	0.96	0.26	0.24	0.35	0.39	0.95	0.88
Codon (M0)	0.83	0.88	0.89	0.72	1.00	0.92	0.35	0.64	0.48	0.22	0.97	0.04

Note: The AA-GTR+G model used the GTR exchangeability rates for amino acids estimated from one of the simulated replicates using codeml (Supplementary Fig. S2). The other 6 data-analysis strategies are defined in the legend to Figure 4, and the results are plotted in Figures 4 and 5 (homo).