Performance of genotypic algorithms for predicting tropism for HIV-1 CRF01_AE recombinant

There is no consensus about the performances of genotypic rules for predicting HIV-1 non-B subtype tropism. Three genotypic methods were compared for CRF01_AE HIV-1 tropism determination.

The V3 env region of 207 HIV-1 CRF01_AE and 178 B subtypes from 17 centers in France and 1 center in Switzerland was sequenced. Tropism was determined by Geno2Pheno algorithm with false positive rate (FPR) 5% or 10%, the 11/25 rule or the combined criteria of the 11/25, net charge rule and NXT/S mutations.

Overall, 72.5%, 59.4%, 86.0%, 90.8% of the 207 HIV-1 CRF01_AE were R5-tropic viruses determined by Geno2pheno FPR5%, Geno2pheno FPR10%, the combined criteria and the 11/25 rule, respectively. A concordance of 82.6% was observed between Geno2pheno FPR5% and the combined criteria for CRF01_AE. The results were nearly similar for the comparison between Geno2pheno FPR5% and the 11/25 rule. More mismatches were observed when Geno2pheno was used with the FPR10%. Neither HIV viral load, nor current or nadir CD4 was associated with the discordance rate between the different algorithms.

Geno2pheno predicted more X4-tropic viruses for this set of CRF01_AE sequences than the combined criteria or the 11/25 rule alone. For a conservative approach, Geno2pheno FPR5% seems to be a good compromise to predict CRF01_AE tropism.