Figure 1.
Phylogenetic relationship between Gag amino acid sequences of VLP panel members (red) and Los Alamos National Laboratory (LANL) subtype reference sequences (black).
Gag subtype reference sequences were downloaded from the LANL website http://www.hiv.lanl.gov/content/sequence/NEWALIGN/align.html and filtered for subtypes present in the VLP panel. The phylogenetic tree was constructed using the neighbour joining method (Clustal W). The scale bar indicates branch length, expressed as the number of substitutions per site.
Table 1.
Amino acid sequence divergence (%) of p24 in VLP panel and LANL reference subtypes after pairwise sequence alignment.
Figure 2.
Total number of VLPs detected per input concentration for each 4th generation combo test and Ag-only test.
Three input concentrations (see legend) of VLPs diluted in negative human plasma were analysed blindly and VLPs were scored “detected” as per the manufacturers instruction for unequivocally positive samples. The left y-axis shows numbers of VLPs detected while the right y-axis shows the corresponding percentage. Dotted lines indicate 50% and 100% detection.
Figure 3.
Overview of all results per test and VLP.
Numbers indicate how many input concentrations were detected per VLP and test, i.e. 3 = 50, 10 and 2 IU/ml concentrations detected, 2 = 50 and 10 IU/ml detected, 1 = 50 IU/ml detected, 0 = VLP not detected at any concentration. The overall sensitivity for each test was calculated as number (#) of VLPs detected (D)/total number of VLPs (n = 129, i.e. # of VLPs detected + # of VLPs not detected [ND], excluding the WHO p24 standard).
Figure 4.
Detection of VLPs before and after heat-denaturation at 10 IU/ml.
VLP preparations of 50 IU/ml were diluted at 1:5 in PBS and heat-denatured for 5 min at 100°C. Results for the non-heat treated VLPs for the Abbot Architect and Perkin Elmer Alliance were taken from the complete panel analysis and heat-treated VLP measurement was conducted separately. For the bioMérieux VIDAS DuoUltra, heat and non-heat treated samples were analysed in parallel. Highlighted in red are VLPs with loss of p24 detection between 1.9–3-fold for the Perkin Elmer Alliance and ≥3-fold for bioMérieux VIDAS DuoUltra and Abbott Architect.