A new ensemble coevolution system for detecting HIV-1 protein coevolution

Li, Guangdi; Theys, Kristof; Verheyen, Jens; Pineda-Peña, Andrea-Clemencia; Khouri, Ricardo; Piampongsant, Supinya; Eusébio, Mónica; Ramon, Jan; Vandamme, Anne-Mieke

doi:10.1186/s13062-014-0031-8

Biology Direct

Table 2 Performance of sequence-based methods in detecting HIV-1 protein coevolution

From: A new ensemble coevolution system for detecting HIV-1 protein coevolution

Method	Area-under-curve (AUC)							Average accuracy		Average harmonic distance		Average Euclidean distance
Method	MA	CA	NC	p6	PR	p6-PR	CSM-PR	L/2	L	L/2	L	L/2	L
APC	0.57	0.55	0.59	0.71	0.57	0.62	0.66	10.8%	8.6%	0.039	0.027	17.38	18.6
ASC	0.56	0.53	0.59	0.75	0.59	0.63	0.62	15.0%	11.7%	0.051	0.028	16.41	18.69
BN	0.71	0.55	0.62	0.69	0.75	0.54	-*	5.9%	5.2%	0.009	0.008	19.94	20.13
CMPro	0.75	0.66	0.85 ^#	0.76	0.74	0.68	0.72	28.9%	22.5%	0.166	0.13	10.05	11.77
CTMP	0.54	0.52	-	-	0.57	0.69	-	3.3%	3.3%	0.004	0.004	16.98	16.98
CoMap	0.52	0.52	0.61	-	0.55	-	0.5	3.9%	4.3%	0.029	0.029	16.85	17.14
Complementary	0.52	0.52	0.57	0.54	0.53	0.53	0.55	4.0%	4.7%	0.008	0.003	19.08	20.01
DCA	0.55	0.55	0.59	0.78	0.51	0.64	0.67	9.2%	7.1%	0.03	0.023	17.43	18.45
DNcon	0.5	0.51	0.66	-	0.61	-	0.77	16.5%	11.3%	0.093	0.07	13.66	15.11
GREMLIN	0.56	0.54	0.6	0.81	0.6	0.6	0.63	13.8%	9.5%	0.04	0.024	17.14	18.77
Interdependency	0.63	0.58	0.68	-	0.66	-	-	7.3%	7.0%	0.028	0.026	18.4	18.58
LogR	0.55	0.54	0.54	0.8	0.55	0.58	0.55	11.4%	8.3%	0.024	0.015	18.44	19.32
MI	0.51	0.54	0.58	0.84	0.58	0.81	0.79	17.9%	12.6%	0.043	0.026	17.6	18.96
MIBP	0.57	0.5	0.57	0.67	0.53	0.62	0.7	4.5%	5.3%	0.021	0.023	17.8	18.12
Mutagenetic	0.53	0.66	0.71	-	0.64	0.86	0.6	15.9%	15.9%	0.027	0.027	19.13	19.13
NBZPX2	0.56	0.52	0.54	0.55	0.54	0.51	0.5	6.1%	5.2%	0.011	0.005	19.51	20.2
NCPS	0.58	0.51	0.54	0.83	0.56	0.86	0.83	17.0%	11.6%	0.018	0.011	19.37	20.27
NNcon	0.68	0.72	0.78	-	0.78	-	-	28.6%	23.8%	0.148	0.132	11.25	12.01
PCC	0.53	0.56	0.55	-	0.51	0.54	0.61	7.0%	5.0%	0.013	0	18.63	20.2
PSICOV	0.56	0.58	0.54	0.55	0.51	0.51	0.53	8.4%	6.2%	0.016	0.012	18.63	18.79
PhyCMAP	0.76	0.7	0.72	0.65	0.72	0.8	0.55	19.4%	17.2%	0.118	0.107	11.83	12.55
PhysicoMI	0.61	0.56	0.52	0.84	0.5	0.72	0.64	7.1%	4.6%	0.009	−0.001	20.46	21.06
RCW	0.54	0.53	0.58	0.82	0.56	0.8	0.78	12.3%	10.9%	0.044	0.032	16.88	18.12
SCA	0.54	0.54	0.56	0.53	0.58	0.77	0.77	15.7%	10.8%	0.027	0.016	18.26	19.08
SVMcon	0.71	0.73	0.67	-	0.77	-	-	24.6%	18.3%	0.14	0.111	11.42	12.65
Spidermonkey	0.58	0.55	0.63	0.67	0.52	0.51	0.57	6.5%	5.7%	0.018	0.01	18.89	19.77
ZRES	0.56	0.53	0.59	0.73	0.56	0.61	0.68	12%	10.7%	0.046	0.032	16.65	18.08
CNPR	0.75(2.5)	0.7(3.5)	0.83(2)	0.84(1)	0.77(2.5)	0.87(1)	0.88(1)	34.7%(1)	26.9%(1)	0.155(2)	0.132(1.5)	10.14(2)	11.52(1)

*AUC was not evaluated due to the lack of long-range couplings predicted. ^#For each column, the numbers in bold indicate methods with the best score among the 28 methods. The ranking of CNPR for each dataset is provided in brackets (see others in Additional file 2: Table S3). Ranking numbers in decimals are results from the average rankings (see examples in Additional file 2: Table S3). Four statistical measurements (AUC, accuracy, harmonic distance, Euclidean distance) are defined in Methods. For the latter 3 methods, the L or L/2 top-ranked predictions were compared and the average scores over the 7 HIV-1 datasets were listed (see performance evaluation per method per dataset in Additional file 2: Table S4-S6).

Back to article page

ISSN: 1745-6150

Contact us

General enquiries: journalsubmissions@springernature.com