Figure 1

Accuracy of reconstructed haplotypes from simulated HIV-1 quasispecies (LQ4_1). We simulated HIV-1 population (LQ4_1; Table 2) with 9 quasispecies consisting of 50, 24, 8, 4, 2, 2, 2, 2 and 1, sequences respectively. The genome sequences within each quasispecies contained at most 0.1 % nucleotide diversity (see Methods). (A) Seven reconstructed whole genome haplotypes (red) were aligned with the 95 simulated genomes. Dark lines represent the branches of the phylogenetic tree (the dotted lines pointing outwards provide better visualization). (B) Average Hamming distance within each quasispecies (black), and between the reconstructed haplotype and the quasispecies at minimum Hamming distance (red). Because one haplotype can represent multiple quasispecies, we used the phylogenetic tree to determine the genomes represented by a haplotype. Markers represent the mean Hamming distance and error bars represent 1 standard deviation. (C) The haplotype frequencies (vertical axis) predicted from haplotype reconstruction pipeline are plotted against the true quasispecies frequencies (horizontal axis). The dominant quasispecies was reconstructed as two haplotypes (Haplo_0 and Haplo_4) and summation of their predicted frequencies is indicated by an open square. The haplotype Haplo_2 (predicted frequency of 7.55%) represents four low prevalence quasispecies: one quasispecies with true frequency of 1.05% (orange circle) and three with true frequencies of 2.1% (see the overlapping points next to orange circle). The points lying on the dashed diagonal line represent perfect haplotype frequency predictions.