Fig 1.
Phylogeny of pneumococcal population structure.
Maximum likelihood phylogeny of 937 pneumococcal carriage isolates inferred from an alignment of 1,111 core COGs including 78,525 polymorphic sites using RAxML with GTR+Γ nucleotide substitution model and 100 bootstrap replicates. Clades are colored by sequence cluster (SC), which are labeled on the outside ring. Some SCs are further divided into monophyletic sub-clusters (A, B, C) based on ancestral recombination history. Pneumococcal serotypes and MLST are labeled on each clade and bootstrap values are labeled on internal branches.
Table 1.
Demography and vaccine type composition of pneumococcal sequence clusters.
For each sequence cluster (SC) or sub-cluster, the number of isolates, proportion of PCV7 and PCV13 vaccine types, and recombination rate (r/m) are listed. The best fit demographic and molecular clock model as determined through comparison of marginal likelihood estimates using BEAST are specified. The Ne directionality (constant, exponentially increasing or decreasing) were determined by assessing the slope of the Ne during the study period (1998–2012) and 95% Highest Posterior Density (HPD).
Fig 2.
Pneumococcal population dynamics pre- and post-vaccine, 1998–2012.
The study periods were subdivided into three epochs and six sub-epochs: Pre-PCV7 [Epoch 1: A (n = 105), B (n = 169)], Post-PCV7 [Epoch 2: A (n = 79), B (n = 319)], and PCV13-Intermediate [Epoch 3: A (n = 119), B (n = 146)]. A.) The proportions of vaccine types (VT) for each epoch from three parent studies (violin) and current study subsample (asterisk). Parent study VT proportions are estimated from serotypes of 3,868 carriage events from previous N/WMA studies [1,6,30]. PCV7 VT include serotypes 4, 6B, 9V, 14, 18C, 19F and 23F. PCV13 vaccine types (minus PCV7 types) include only serotypes 1, 3, 5, 6A, 7F, 19A. Violin plots represent the realization of 1000 bootstrap replicates subsampling with replacement from each epoch. Asterisks represent the point estimates for VT proportions in the current study. B.) Simpson’s Diversity Index of serotype diversity across study periods estimated from three parent studies (n = 3,868). This measure summarizes the number and abundance of each serotype. C.) Population genetic measures of diversity, Watterson estimator (Θw) and π (average number of pairwise differences), estimated from 4-fold (synonymous) degenerate sites of taxa in sub-epochs of the current study. D.) Population genetic statistic Tajima’s D estimated from the core genomes of taxa in sub-epochs of the current study. Negative values of Tajima’s D indicate many sites with a rare minor alleles.
Fig 3.
Population structure of pneumococcal carriage isolates among N/WMA.
Populations are divided into sequence clusters (SC) based on genomic data and further subdivided into three epochs based on collection date. The bars represent the proportion of population comprising each SC during an epoch and are stratified by serotype composition. Solid bars represent non-vaccine serotypes; checkered hatched pattern PCV7 vaccine types; and vertical line pattern PCV13 vaccine types (those not included in PCV7). Serotypes comprising each SC are also labeled above each column. All SCs except SC4 and SC27 are monophyletic. SC4 is has three distinct sub-clades and SC27 include polyphyletic lineages present at minor frequencies in the population.
Fig 4.
Mean squared errors (MSEs) comparing changes in genomic loci frequencies between Epoch 1A and all subsequent sub-epochs.
For each comparison, 75 individuals were subsampled with replacement from each sub-epoch and 1,000 bootstrap replicates were performed. A.) MSEs for sub-epoch comparison of frequencies of 22,434 biallelic SNP sites found among 256 metabolic genes. B.) MSEs for sub-epoch comparison of frequencies of 53 variants of 19 polymorphic antigens. C.) MSEs for sub-epoch comparison of frequencies of 2370 COGs found from 5–95% among all 937 taxa.
Fig 5.
Association of hierBAPS sequence cluster (SC) with polymorphic non-capsular antigens.
Maximum likelihood phylogeny of 937 pneumococcal carriage isolates corresponding to Fig 3. Color ramps in legend designate SCs. Antigens include pspA, SP2194, SP0609, phtD, ply, pspC, rrgA/B/C (type 1 pilus islet), stkP, strH, nanA, zmpA, cbpE, hyl, eno, pepO, pavA, and GroEL. Accession numbers for antigen variants have been previously published [15].
Fig 6.
Bar plots comparing proportion of variants among 19 polymorphic antigens between Epochs E1-E3.
Each plot represents the change in distribution of polymorphic antigen variants during Epoch 1 (E1), Epoch 2 (E2), and Epoch 3 (E3). Antigens are labeled above each bar plot, and variants of each antigen are colored according to the legend below each plot. Antigen labels are annotated to indicate whether the change in the distribution of frequencies was significant between E1 and E2 (“1”), E2 and E3 (“2”), or both (“12”). All antigens are found among all strains with the exception of rrgA/B/C, which are only present in a subset of taxa. Accession numbers for each variant has been previously published.