Related Experiment Video
Updated: Jun 4, 2026
Genotypic Inference of HIV-1 Tropism Using Population-based Sequencing of V3
Published on: December 27, 2010
[Variability in HIV viral tropism determination using different genotypic algorithms in patients infected with B
José Javier Rodríguez1, Eduardo Seclén, Eva Poveda
1Servicio de Microbiología, Hospital de Conxo-CHUS, Santiago de Compostela, La Coruña, España.
Background:
Genotypic tools based on the analysis of the V3 region are seen as an alternative to phenotypic assays for viral tropism determination before prescribing maraviroc. The concordance between different genotypic algorithms has been evaluated in HIV+ patients infected with B versus non-B subtypes.
Methods:
HIV-infected patients on regular follow up at Hospital Universitario de Santiago de Compostela (Spain) were selected. The env-V3 region was sequenced from plasma samples and viral tropism was estimated using 8 different genotypic algorithms. Concordance among predictors was statistically evaluated by the calculation of the kappa index. Phylogenetic analyses were performed to determine the genetic subtype.
Results:
A total of 92 HIV-infected patients were selected, 72 B and 20 non-B subtypes. Regarding the B subtype group, significant kappa values were obtained among all 28 possible combinations between the genotypic predictors evaluated. The best concordance among non-related predictors was observed for webPSSM(SINSI)/Wetcat(PART) (k: 0.771) and webPSSM(SINSI)/geno2pheno (k: 0.574). Conversely, among non-B subtypes, a significative kappa index was only obtained for 13 combinations. Among non-B subtypes, the best concordance values were obtained for webPSSM(X4R5)/Wetcat(PART) (k: 0.600) and webPSSM(SINSI)/Charge rule (k: 0.590).
Conclusion:
A high concordance was observed between different genotypic algorithms to determine viral tropism among HIV-1 B subtypes infected patients, especially between webPSSM(SINSI) and geno2pheno or Wetcat. Conversely, the overall concordance among non-B subtypes was lower. This heterogeneity could be justified by the low prevalence of non B subtypes in the datasets in which the genotypic tropism predictors were trained.
