Evaluation of genotypic prediction of HIV-1 tropism using population sequencing of replicates.

Determination of human immunodeficiency virus tropism has contributed to the understanding of the pathogenesis of HIV and is necessary prior to the use of CCR5 antagonists. Replicate V3 sequences may generate different sequences and improve viral tropism prediction. The diversity of HIV was evaluated to access its influence on prediction. Plasma RNA was retro-transcribed and amplified using a one-step protocol, followed by nested PCR and sequencing using an ABI3130XL. Eighty-one patients, 74% male and 26% female, with a median age of 44 years had either a single sequence (n=50) or 2-4 replicates (n=31) evaluated. Most patients (92%) had used multiple anti-retroviral regimens. Tropism prediction was performed using the Geno2pheno clonal option. The number of ambiguous nucleotides, the deduced non-synonymous amino acids at V3 and the genetic distance were quantified. Using a 20% false positive rate (FPR) cut-off, 41/81 (50.6%) was predicted as X4. TCD4 was lower, 226 cells/mm(3) (IQR 82-378), in patients infected with X4; TCD4 for R5 was 324 cells/mm(3) (IQR 200-538, p<0.05). The number of ambiguous nucleotides correlated with a lower FPR value (p<0.0027). Although different sequences may be generated, the number of replicates was not associated to a lower FPR or X4 assignment, and may allow a better prediction of this biological characteristic. Ambiguous nucleotides correlate inversely to a lower FPR.