Computational protocol: HIV-1 drug resistance mutations emerging on darunavir therapy in PI-naive and -experienced patients in the UK

Similar protocols

Protocol publication

[…] UK CHIC participants (all aged over 16 years) were eligible for the study if they had received at least 30 days of a darunavir-containing regimen and had both a ‘baseline’ (defined as any time prior to darunavir exposure) and ‘post-exposure’ genotypic resistance test result (obtained either during darunavir therapy or within 30 days of stopping this agent). Participants were excluded if they had received another PI for ≥90 days between the baseline and post-exposure tests, to avoid attributing the effect of other agents to darunavir. A 90 day period was allowed to enable a resistance test result to be obtained and acted upon prior to switching to darunavir from another PI. Only genotypes with a complete protease sequence were considered. If more than one baseline genotypic test had been performed, the one closest to the start of darunavir was used. If more than one post-exposure test had been performed, the results were combined and therefore reflect cumulative resistance. All ART regimens that included darunavir were considered. Information on the dosing frequency (once or twice daily) was not recorded. The prevalence at baseline of darunavir DRMs, defined according to the IAS-USA 2015 list, was assessed from the baseline protease sequences. Emergent DRMs were identified by comparing the baseline and post-exposure sequences for each individual. Viral subtypes were determined by analysing the pol sequence with the REGA HIV subtyping tool version 3., All analyses were stratified by history of exposure to other PIs prior to initiating darunavir. Statistical analyses were performed with Stata/IC 13.1 software (StataCorp LP, College Station, TX, USA).Selection pressure was examined by estimating non-synonymous (dN) and synonymous (dS) substitution rates during darunavir therapy using the HyPhy software package available on Datamonkey, a web-based interface., The dN:dS ratio was calculated for amino acid sites 5–99 of the protease gene and positive selection was inferred if dN > dS. The analysis was restricted to those with subtype B infection to avoid introducing bias from inter-subtype variability in WT amino acids and polymorphic loci. A case control approach was used for the positive selection analysis. Cases met the study eligibility criteria above and additionally had not been exposed to any other PIs prior to initiating darunavir. To distinguish the effects of darunavir therapy selection pressure from the natural evolution of the protease gene over time, we selected controls who met the same inclusion criteria, but who were PI-naive and had initiated an NNRTI-based ART regimen. For each case, two controls were randomly chosen, matched by calendar year of initiation of either darunavir or NNRTI. The selection pressure analysis was performed using three different codon-based algorithms: fixed effects likelihood (FEL), single likelihood ancestor counting (SLAC) and fast unconstrained Bayesian approximation (FUBAR) using a cut-off P value <0.05 for FEL and SLAC, and a posterior probability >0.95 for FUBAR., A starting phylogenetic tree was supplied for each analysis that was inferred by maximum-likelihood method in FastTree using a general time reversible nucleotide model of substitution. A single breakpoint recombination tree was used if recombination was detected. Sites were considered to be positively selected if this was confirmed by at least two of the three algorithms. […]

Pipeline specifications

Software tools REGA HIV Subtyping Tool, HyPhy, Datamonkey, FastTree
Applications Phylogenetics, Population genetic analysis
Organisms Human immunodeficiency virus 1