Frequent CXCR4 tropism of HIV-1 subtype A and CRF02_AG during late-stage disease - indication of an evolving epidemic in West Africa

Table 3 Comparison of different genotypic rules and bioinformatic tools for prediction of HIV-1 coreceptor tropism based on the HIV-1 V3 amino acid sequence.

		No. patients with virus phenotype:		Performance (%)
Prediction method	Predicted phenotype	R5	R5X4/X4	Sensitivity	Specificity	PPV⁴	NPV⁵
11/25	R5	71	25	31	95	73	74
	X4	4	11
Net ≥ +5	R5	65	20	44	87	62	76
	X4	10	16
Raymond et al.¹	R5	73	22	39	97	88	77
	X4	2	14
WebPSSM_X4R5²	R5	63	22	39	84	54	74
	X4	12	14
WebPSSM_SINSI²	R5	69	25	31	92	65	73
	X4	6	11
Geno2Pheno³	R5	66	18	50	88	67	79
	X4	9	18
Net ≥ +4	R5	35	10	72	47	46	78
	X4	30	26
Total ≥ 8	R5	63	18	50	84	60	78
	X4	12	18
Net ≥ +5 and Total ≥ 8	R5	55	13	64	73	53	81
	X4	20	23

¹Raymond et al. used a combined rule were one of the following criteria was required for HIV-1 CRF02_AG CXCR4-tropism: (i) R or K at position 11 of V3 and/or K at position 25, (ii) R at position 25 of V3 and a net charge of ≥ + 5, or (iii) a net charge of ≥ + 6.
²Position-specific scoring matrix, http://indra.mullins.microbiol.washington.edu/webpssm/ (January 2010).
³Geno2Pheno was used with a false-positive rate of 10%. http://coreceptor.bioinf.mpi-inf.mpg.de/ (January 2010).
⁴Positive predictive value.
⁵Negative predictive value.

ISSN: 1742-4690