Fig 1.
A flow chart describing the tree and network construction process.
Fig 2.
(A) A contrived evolution of a mixed viral population involving five mutations and six clones. Dotted lines indicate internal nodes extended to a leaf. (B) A notional representation of sequencing across the region of interest, and the resulting Depth-Position graph. Paired reads bridge two clusters of mutations. (C) Read count data obtained for the two clusters, with total depth x1000, along with artifacts †. (D) Evolutionary trees corresponding to (Ci,ii). (E) Ordered list of mutations and population prevalences (%). (F) Reconstruction of original tree in (A).
Fig 3.
(A) A pedagogic evolution recombination network with four mutations and two recombination events across two segments. (B) Typical haplotype tables arising from (A). (Ci) Two clusters of mutations grouped by paired reads on two segments. (Cii) Clones C1000 and C0010 undergo within segment recombination into C1010, with a crossover site between M1 and M3. (Ciii) Clones C1001 and C1110 undergo between segment recombination (reassortment) into C1111. (D) Recombination networks arising from the haplotype tables in (B). (E) The prevalence of the four mutations across five days. (F) Phylogenetic network associated with (A). (G) Point and range prevalence estimates. (Hi) A network consistent with the two networks of (D). (Hii) Incompatible prevalence conditions associated with (Hi).
Fig 4.
(A) Size and function of the eight flu segments. (B,C) Depth of mutations from segment 4 from host 2761 Day 4 and host 6292 Day 3, respectively. (i,ii) Results from Hi-Seq and Mi-Seq experiments on separate libraries from the same samples.
Fig 5.
(A) Evolution tree arising from five mutations on segment 6. (B) Prevalences of the clones across the times series.
Fig 6.
(A) Three template switching models. (B) Fitted models for three pairs of linked mutations on segment 4.
Fig 7.
(A) Mutation prevalences of six mutations in segment 4. (B) Prevalences of the ten associated clones. (C) Two tables of linked mutations exhibiting network like relationship of mutations 709 and 1401, and tree like relationship of 1387 and 1401. (D) The phylogenetic network of the single fitted cloneset. (E) One of 22 possible recombination networks that arise from (D). (F) Probable tree structure from (E) after template switching is considered.
Fig 8.
Three mutations between three segments that indicate two re-assortment events.
(A) Mutation prevalences across time series. (B) One of 51 recombination networks that fit the data. (C) Two phylogenetic networks that fit the data ((i) and (ii)-(iv)), corresponding to four clonesets. (D) Prevalence ranges for the four clonesets.
Fig 9.
(A) Cayley trees for 2, 3 and 4 vertices. (Bi) Vertex list v for example (*). (Bii) Prüfer sequence p. (Biii) Tree construction. (Ci) The graph directed away from the root. (ii) The equivalent compact clonal tree. (iii) The corresponding complete clonal tree. (D) Alignment of trees * and † to haplotype tables.
Fig 10.
Validation profiles for a range of haplotype counts, including; (A) Recall, (B) Precision, and (C) Error.
In all cases the solid line denotes the algorithm preformance, the dashed line indicates Shorah performance.