Identify High-Quality Protein Structural Models by Enhanced -Means

Table 1

Comparison between -means, -means++, and SPICKER on 56 protein decoys.

Index	PDB	Len	Size	Best	-means++	-means	-means	SPICKER	Random
1	1abv	103	526	0.507	0.3701	0.3834	0.4910	0.3813	0.479
2	1af7	72	527	0.623	0.5009	0.5009	0.4820	0.4874	0.322
3	1ah9	63	510	0.696	0.5040	0.4743	0.4740	0.4657	0.434
4	1aoy	65	529	0.711	0.6482	0.6695	0.6695	0.6695	0.622
5	1b4bA	71	460	0.473	0.3815	0.4279	0.4270	0.4501	0.379
6	1b72A	49	534	0.697	0.5397	0.3917	0.6410	0.4923	0.562
7	1bm8	99	329	0.388	0.4332	0.3787	0.3320	0.3550	0.255
8	1bq9A	53	573	0.465	0.3540	0.3459	0.3990	0.3873	0.411
9	1cewI	108	452	0.748	0.7294	0.7154	0.7290	0.7187	0.617
10	1cqkA	101	284	0.885	0.8439	0.8539	0.8539	0.8539	0.815
11	1csp	67	315	0.753	0.7158	0.7158	0.7158	0.7158	0.686
12	1cy5A	92	273	0.893	0.8685	0.8839	0.8680	0.8839	0.876
13	1dcjA	73	525	0.368	0.3299	0.3645	0.3170	0.3264	0.334
14	1di2A	69	374	0.843	0.7622	0.7663	0.7620	0.7663	0.374
15	1dtjA	74	285	0.814	0.7901	0.7581	0.7370	0.7901	0.705
16	1egxA	115	352	0.827	0.7673	0.7673	0.7673	0.7673	0.768
17	1fadA	92	514	0.652	0.5716	0.5755	0.5755	0.5755	0.553
18	1fo5A	85	340	0.568	0.5391	0.5391	0.5230	0.5296	0.469
19	1g1cA	98	307	0.787	0.7473	0.7732	0.7800	0.7732	0.621
20	1gjxA	77	525	0.515	0.2375	0.3807	0.3810	0.4298	0.191
21	1gnuA	117	553	0.647	0.5353	0.5353	0.5350	0.5456	0.509
22	1gpt	47	469	0.553	0.5130	0.5377	0.5060	0.4927	0.517
23	1gyvA	117	337	0.776	0.7406	0.7406	0.7540	0.7406	0.753
24	1hbkA	89	300	0.708	0.6633	0.6633	0.6633	0.6633	0.599
25	1itpA	68	526	0.511	0.3069	0.3152	0.3150	0.3096	0.335
26	1jnuA	104	269	0.768	0.7457	0.7237	0.6980	0.7237	0.711
27	1kjs	74	548	0.5	0.3728	0.3728	0.3580	0.3728	0.313
28	1kviA	68	550	0.79	0.7181	0.6774	0.7220	0.6774	0.642
29	1mkyA3	81	285	0.552	0.4155	0.4155	0.4155	0.4155	0.384
30	1mla_2	70	335	0.775	0.6742	0.6226	0.6226	0.6226	0.609
31	1mn8A	84	545	0.457	0.2517	0.3543	0.3540	0.3285	0.310
32	1n0uA4	69	301	0.588	0.4753	0.4746	0.4524	0.4524	0.333
33	1ne3A	56	566	0.453	0.2523	0.3943	0.3940	0.3724	0.344
34	1no5A	93	426	0.419	0.3710	0.4247	0.4240	0.4054	0.500
35	1npsA	88	469	0.800	0.7671	0.7671	0.2810	0.7671	0.283
36	1o2fB	77	510	0.528	0.3380	0.338	0.3370	0.2690	0.379
37	1of9A	77	507	0.585	0.5469	0.494	0.5460	0.4940	0.554
38	1ogwA	72	520	0.890	0.7853	0.7853	0.7850	0.8622	0.78
39	1orgA	118	442	0.816	0.7440	0.7339	0.7440	0.7440	0.693
40	1pgx	59	562	0.551	0.5824	0.3216	0.5160	0.4446	0.51
41	1r69	61	291	0.824	0.7007	0.7255	0.7255	0.7255	0.827
42	1sfp	111	308	0.758	0.7453	0.7453	0.7454	0.7454	0.749
43	1shfA	59	536	0.836	0.5649	0.5070	0.5640	0.5070	0.408
44	1sro	71	515	0.648	0.6513	0.6513	0.5820	0.6158	0.583
45	1ten	87	294	0.851	0.8215	0.8215	0.7860	0.8215	0.781
46	1tfi	47	339	0.592	0.5061	0.5576	0.5520	0.5576	0.550
47	1thx	108	302	0.865	0.8000	0.8000	0.8000	0.8000	0.819
48	1tif	59	542	0.340	0.3269	0.2667	0.2660	0.3199	0.232
49	1tig	88	565	0.585	0.5524	0.4596	0.4740	0.4176	0.517
50	1vcc	76	551	0.455	0.3973	0.4066	0.3970	0.4066	0.291
51	256bA	106	506	0.814	0.7657	0.7578	0.7650	0.7578	0.723
52	2a0b	118	282	0.838	0.8083	0.8083	0.8083	0.8083	0.768
53	2cr7A	60	540	0.666	0.3589	0.5059	0.5820	0.5136	0.365
54	2f3nA	65	485	0.758	0.6403	0.7322	0.6510	0.7132	0.626
55	2pcy	99	435	0.637	0.6040	0.5795	0.6460	0.6233	0.527
56	2reb_2	60	550	0.403	0.3902	0.378	0.3290	0.3174	0.416

The length of the protein sequence.
The size of the models in the decoy.
The best (maximum) TM-score of the models in the decoy.
The TM-score of centroid model in the largest cluster selected by -means++ (bold indicates better than SPICKER).
The TM-score of centroid model in the largest cluster selected by -means (bold indicates better than SPICKER).
The TM-score of centroid model in the largest cluster selected by -means (bold indicates better than SPICKER).
The TM-score of centroid model in the largest cluster selected by SPICKER.
The TM-score of centroid model selected by random.