Table 1.
Details of samples sequenced from clinical virus samples.
Fig 1.
Nucleotide diversity π for acute infections across different virus samples and genes.
(A) Each point represents the π diversity of a single sample, across all genes sequenced. Diversity values were calculated using transition mutations only. (B) Gene by gene breakdown of nucleotide π diversity. “gag” represent the gag-pol reading frame. Values for HIV envelope (env) (squares) were taken from previously published data [15].
Fig 2.
Variant frequency plots in representative samples.
Shown are frequencies of transition variants called by AccuNGS, for representative samples from each virus (HIV, top row, RSV, middle row, CMV, bottom row). Variant frequencies lower than 1% are faded. Samples exemplify mixed genotype infections (HIV6, CMV2), mutation biases and presumable hypermutation via host editing (HIV9, RSV samples), and relatively homogenous populations (HIV3, CMV5, CMV4) (see text for details).
Fig 3.
Illustration of method for haplotype reconstruction.
The method searches for enrichment of pairs of mutations on the same read, and concatenation of enriched reads that share a mutation into a reconstructed minor haplotype. Notably, the concatenation approach is suitable for populations with limited diversity, as is the case in acute infections; in highly diverse populations, many haplotypes may share the “blue” mutation illustrated in the figure.
Fig 4.
Haplotype reconstruction based on co-occurrence of variants on the same reads.
Shown are inferred haplotypes (lines) based on consecutive significant associations of pairs of variants (shapes) one to another on the same read. The uppermost line in each panel represents the consensus sequence, which by definition is the major haplotype in each sample. Both HIV6 and CMV2 samples show strong evidence of an additional haplotype, which is likely a second founder genotype. Sample HIV9 shows evidence of G>A hyper-mutation in the context of APOBEC3 editing, samples RSV4 and RSV15 show evidence of T>C or A>G hyper-mutation in the context of ADAR editing in regions spanning a few hundred bases. The hyper-mutated region in RSV4 sample is magnified for clarity. “Empty” panels signify what are likely single haplotype infections, with no evidence of hyper-mutation.