Ethics Statement

The studies providing cervical samples for this work have been IRB approved by each ethics committee and all samples received in the Burk lab were coded and did not have individual identifying information. In details, Rwanda - The Rwanda National Ethics Committee and the Institutional Review Board of Montefiore Medical Center, Bronx NY approved the study protocol and the consent process. Burkina Faso - The Yerelon Cohort Research Programme was approved by the Ethical Committee of the Centre Muraz, Bobo Dioulasso, the National Ethical Committee of the Ministry of Health, Burkina Faso, and the Ethics Committee of the London School of Hygiene and Tropical Medicine, Zambia - The study was approved by the Research Ethics Committee of the University of Zambia and the Institutional Review Board of the University of Alabama at Birmingham, Thailand - The study protocols were reviewed and approved by the committees on human subject research at Johns Hopkins Bloomberg School of Public Health, Baltimore, MD; Merck & Co., Inc., West Point, PA, each participating recruitment site, and the Institutional Review Board of the Thailand Ministry of Health (MOH), Thailand. Taiwan - The study was approved by the Institution Review Board of the National Taiwan University University College of Public Health. Guanacaste, Costa Rica - The study and informed consent forms were approved by Institutional Review Boards of Costa Rica and the U.S. National Cancer Institute.

Clinical Specimens, Identification of Novel HPV Variants and Whole Genome Sequencing

DNA from cervicovaginal samples already determined to have alpha-7 types (HPV18, 39, 45, 59, 68 and 70) by previous testing were available from women participating in epidemiological studies worldwide, including - Costa Rica [29], Taiwan [30], Thailand [31,32], Rwanda [33], Burkina Faso [34] and Zambia [35]. The methods for sample collection and HPV typing are provided in the references from each study. The number of samples analyzed for each type is shown in Table 1. The HPV genomes within the DNA samples were classified by sequencing the URR and/or E6 regions from PCR products as described [12,27]. Briefly, we used type-specific primers to amplify a partial fragment of the URR region and/or the E6 ORF using a one-tube nested PCR method [36]. The E6 ORF was evaluated only for those specimens that did not yield data for the URR region. The PCR product sizes were confirmed by gel electrophoresis, purified using the QuickStep 2 PCR Purification kit (Edge BioSystems, Gaithersburg, MD) or QIAquick Gel Extraction kit (Qiagen, Valencia, CA) and submitted for Sanger sequencing of both strands at the Einstein Genomics Facility. The sequences for each type were separately aligned and preliminary phylogenetic trees were used to identify samples that likely contained divergent viral genomes (data not shown). Based on this analysis, we selected type-specific viral isolates for complete genome sequencing that (1) represented novel variant clades or (2) had 2 or more isolates that contained at least 2 unique sequence variations (e.g., single nucleotide polymorphisms (SNPs)) not present in other isolates within the URR/E6 regions.

The number of genomes selected for sequencing was based on identification of divergent isolates as described above and differed for each type (Table 1). The complete 8 kb genomes from clinical samples were amplified in 2 to 3 overlapping fragments using type-specific primer sets (available from authors) based on the prototype sequence of each type. For overlapping PCR, an equal mixture of AmpliTaq Gold DNA polymerase (Applied Biosystems, Carlsbad, CA) and Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA) was utilized as previously described [12,37]. PCR products of anticipated size, as determined by gel analyses, were either directly sequenced or cloned into pGEM-T easy (Promega, Madison, WI) or TOPO TA pCR2.1 vectors (Invitrogen, Carlsbad, CA) and then sequenced. Comparison of repeat sequencing of PCR products from the same isolates resulted in a difference of less than one change per 8,000 bp; whereas, comparison of the cloned genomes gave a difference of approximately one difference per 5,000 bp. For discrepancies between sequences, we used the sequence of the PCR product as the consensus sequence. HPV complete genome sequences were submitted to GenBank (Table S1). In addition, a GenBank search, at the time of initial analysis (November 2012), for alpha-7 complete genomes identified 9 HPV18 isolates from Thailand [38], 1 HPV59 and 2 HPV68 isolates from China [39,40] and the set of reference sequences (i.e., HPV18, HPV39, HPV45, HPV59, HPV68 and HPV70.) ME180, previously a subtype of HPV68, was included as a variant of HPV68. HPV85 and HPV97 genomes were not investigated in this study due to the limited diversity of currently reported genomes [18]. The accession numbers of all sequences used in this report are listed in Table S1.

Evolutionary analyses and phylogenetic tree construction

The nucleotide sequences of the complete circular genomes were linearized at the first ATG of the E1 ORF and globally aligned using the program MAFFT v 6.864b [41]. Based on the concept of a single ancestor for each type, a unique genome size is assigned to each HPV type based on the global alignment; the variation in genome sizes of isolated variants is the result of insertions and deletions (indels). Each indel was counted as one event. The assignment of position numbers for each nucleotide is based on the nucleotide numbering of the prototype reference sequence.

Maximum likelihood (ML) trees were constructed using RAxML MPI v 7.2.8.27 [42] and PhyML MPI v 1.4.3 [43] with optimized parameters based on the aligned complete genome nucleotide sequences. MrBayes v 3.1.2 [44] with 10,000,000 cycles for the Markov chain Monte Carlo (MCMC) algorithm was used to generate Bayesian trees. A 10% discarded burn-in was set to eliminate iterations at the beginning of the MCMC run. For Bayesian tree construction, the computer program ModelTest v 3.7 [45] was used to identify the best evolutionary model; the identified GTR model was set for among-site rate variation and allowed substitution rates of aligned sequences to be different. Data were bootstrap resampled 1,000 times in PhyML.

SNPs within the HPV genomes and lineage-specific SNPs were determined from alignments of type specific variant genomes using MEGA v 5.05 [46] and MacClade v 4.08 [47], respectively. Mean nucleotide differences and standard errors between and within type-specific lineages and sublineages were calculated from the global sequence alignment of each type using MEGA v 5.05 bootstrapped 1,000 times [46]. The p-distance method in MEGA v 5.05 [46] was used to calculate the genome-genome pairwise differences from the above alignments. The rarefaction curves for each type were generated by EstimateS v 8.2 [48].

HPV variant distribution, lineage classification and nomenclature

HPV isolates for sequencing were selected from a large set of samples previously typed for HPV and further analyzed for novel variants by sequencing a fragment of the URR and/or E6 ORF (Table 1). From the complete genome sequences of the isolates sequenced in this study and those obtained from GenBank, HPV variant lineage classification and nomenclature were assigned, as previously described [10,12]. The phylogenetic topologies for each type combined with an approximate cut off of 1.0% difference between complete genomes were used to define major variant lineages (1.0%-10.0%). Each major lineage was named using an alphanumeric, with the “A” clade always containing the reference genome for each type. Differences between genomes in the 0.5%-1.0% range were designated as sublineages (e.g., A1, A2, etc.). When the reference genome could be classified as a sublineage, it was always assigned into the “A1” clade. An overview of the alpha-7 HPV types and variant lineages are shown in Figure 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Alpha-7 phylogenetic tree showing representative types and variant lineages.

A maximum likelihood (ML) tree was constructed using RAxML MPI v 7.2.8.27 [42] and PhyML MPI v 1.4.3 [43] inferred from the global alignment of complete circular genome nucleotide sequences linearized at the first ATG of the E1 ORF. Numbers on or near branches indicate support indices < 100% by RAxML and PhyML, respectively. The shaded areas represent groupings of lineages and sublineages of HPV18, HPV39, HPV45, HPV59, HPV68 and HPV70; the prototype sequences of HPV85 and HPV97 are also included as indicated, although no variant lineages are distinguished due to the limited number of isolates of these types. The length of broken and solid lines represent distance between clades, although the number of changes is different for these two lines, as indicated in the upper left corner of the figure. HPV51, an alpha-5 type, was set as the outgroup.

https://doi.org/10.1371/journal.pone.0072565.g001

Genomic Diversity of HPV18 Variants

HPV18 complete genome variants have been previously investigated and two major lineages, European (E) and African (Af) were identified [14,18]. The URR and/or E6 regions of 380 HPV18 isolates from cervical samples were sequenced and clustered into distinct groups as shown in Figure 2 [14,18]. Numerous novel SNPs or SNP patterns, e.g., G437A, C7549A and T7475G (Figure S1A), were detected in several samples from which 23 variants with maximum diversity were selected for complete genome sequencing (Table 1). In addition, twenty-two HPV18 complete genome variants from Costa Rican (n = 13) [18] and Thai women (n = 9) [38] that were previously published (Table S1) are included in the current analyses.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. HPV18 tree topology and pairwise comparisons of individual complete genomes.

Phylogenetic trees were inferred from global alignment of complete genome nucleotide sequences (the other alpha-7 HPV reference prototypes were set as the outgroup). Numbers on or near branches indicate support indices in the following order: maximum likelihood (ML) bootstrap percentages using RAxML MPI v 7.2.8.27 [42] and PhyML MPI v 1.4.3 [43], and Bayesian credibility value percentage using MrBayes v 3.1.2 [44]. An asterisk (*) indicates 100% agreement between methods. “NA” indicates disagreement between a method and the reference RAxML tree at a given node. Thus, one tree is shown, but three different methods of tree construction were used to estimate the support of the provided tree, as explained above. Distinct variant lineages (i.e., termed A, B, and C) are classified according to the topology and nucleotide sequence differences from > 1% to < 10%; distinct sublineages (e.g., termed A1 and A2) were also inferred from the tree topology and nucleotide sequence differences in the > 0.5% to < 1% range. The percent nucleotide differences for each isolate compared to all other isolates (i.e., 1 x 1 comparisons) are shown in the panel to the right of the phylogeny. Values for each comparison of a given isolate are connected by lines and the comparison to self is indicated by the 0% difference point. Symbols and colored lines are used to distinguish each isolate.

https://doi.org/10.1371/journal.pone.0072565.g002

A total of 422/7857 (5.4%) nucleotide positions showed variations whereas, 165/2476 (6.7%) codons specified variable amino acids (i.e., they represent nonsynonymous SNPs) (Table 1). The maximum nucleotide pairwise difference between the most dissimilar isolates was 2.1% (Table S2, second column). The most variable region was the noncoding region 2 (NCR2) between the E5 and L2 ORFs with 11.6% overall nucleotide diversity. Consistent with a previous report [18], three different indel events were detected across the genome in the following regions: the E2/E4 ORFs (6 bp deletion), the noncoding region 1 (NCR1 is the region between the stop codon of the E2 ORF and the start codon of the E5 ORF) (19 to 20 bp deletion), and the URR (7 bp deletion) (Figure S1A).

Phylogenetic analyses clustered HPV18 variants into two deeply separated groups (Figure 2). The previous non-Af variants were designated lineage A, and the majority of Af variants were classified as lineage B. A third new lineage was identified based on 2 isolates, Qv39775 and BF226; the complete genome nucleotide sequence differences of lineage C compared to lineage B variants were 1.0%-1.2% (Table S3). Lineage A was relatively distant to lineages B and C with mean differences of 1.92% ± 0.15% and 1.71% ± 0.14%, respectively. Lineage C variants could be distinguished by 23 SNPs throughout the complete genome (E6 - G437A; E1 - T1526C, A1636G; E2 – G3757C, C3790T; E5 – T3942C, T3963G; L2- T4423G, G4486A, A4513T, A4726G, G5191A, C5225A, G5351A; L1 – A5638G, G5764A, A5880G, T5942G, A6224T, T6536C, A6644C, A6845C; and URR- T7475G), while lineages A and B had 62 and 33 lineage specific nucleotide variations, respectively (Figure S1A and Figure S2).

Lineage A variants were further divided into 5 sublineages A1 - A5 with mean differences ranging from 0.37% ± 0.06% to 0.83% ± 0.10% (Figure 2 and Table S3). Previously termed HPV18 Asian-American (E1(AA)) variants clustered into sublineages A1 and A2; sublineages A3 and A4 variants were composed of European (E2) variants [18]. An isolate CU11 from a Thai woman [38] represented sublineage A5 (0.73%-0.83% different to other lineage A variants). The HPV18 prototype (HPV18|Ref) was assigned to the A1 sublineage. Lineage B could be subdivided into three sublineages: B1, B2 and B3 (Figure 2) based on tree topologies; sublineages B1/B2 and B3 have also been referred to as African-1 (Af1) and Af2, respectively [14,18].

The distribution of HPV18 variant lineages was consistent with previous geographic placement [14], however further classification of lineages revealed a few insights. Essentially all the isolates (96%, 24/25) from Taiwan and Thailand mapped to the A1 sublineage. Most isolates from southern-central African countries of Rwanda (62%, 16/26) and Zambia (50%, 6/12) were from the B1 lineage; whereas samples from Burkina Faso in western Africa were classified as B2 (46%, 5/11) or B3 (36%, 4/11). In contrast, the samples from Costa Rica showed a mixed distribution consisting of 45% (134/298) lineage A3, 22% (65/298) lineage B1 and 17% (50/298) lineage A1.

Genomic diversity of HPV39 variants

We amplified and sequenced the E6/URR region of 122 HPV39-containing samples and selected 19 isolates for genome sequencing (Table 1 and Table S1). The overall nucleotide and amino acid variable positions of the complete genomes (19 sequenced for this analysis plus the reference sequence makes a total of 20 complete genomes) were 2.4% (186 sites among 7912 nt) and 2.5% (61 sites among 2428 aa), respectively. The noncoding regions (NCR2 and URR) were more variable than the coding ORFs. HPV39 isolates Tw562 and BF182 were the most distantly related genomes with a nucleotide sequence pairwise difference of 1.1%; this distance represented the maximum inter-lineage diversity of HPV39 variants (Table S2).

Phylogenetic analyses based on complete genome sequences clustered HPV39 variants into 2 lineages designated A and B (Figure 3). Lineage A was further divided into two sublineages, A1 and A2 (0.48 ± 0.06% different); both sublineages were nearly equally distant to lineage B, with mean differences of 0.98 ± 0.09% and 0.97 ± 0.08%, respectively (Figure 3 and Table S3). Isolate Qv36565 (an A1 variant) may bridge the evolution of sublineages A1 and A2, sharing a set of nucleotide changes with A2 variants (E1 - C1298G, C1683T, A2257C, G2355A; E2 – T3157C; E5 – A4081G; L1 – C6785A) (Figure S1B). Lineage and sublineage specific nucleotide variations were determine across the genomes (3 SNPs for sublineage A1, 7 for A2, 37 for lineage A, and 35 for lineage B) (Figure S2). Sublineage A2 variants had a nine amino acid repeat within the E1 ORF (DAEGE(H/N) GGS), and a 52 bp repeat was detected within the URR region of isolate Qv29509 (an A1 variant).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. HPV39 variant tree topology and pairwise comparisons of individual complete genomes.

The phylogenetic tree was constructed as described in Figure 2. Distinct variant lineages and sublineages were determined as described in Figure 2. The percent nucleotide sequence differences are shown in the panel to the right of the phylogenetic tree as described in Figure 2.

https://doi.org/10.1371/journal.pone.0072565.g003

All HPV39 isolates from Costa Rica mapped to either the A1 (67%, 70/105) or A2 (33%, 35/105) lineages; whereas, the few samples from Taiwan sorted to A2 (2/4) or B (2/4), the Rwandan samples sorted to lineages A (7/8) or B (1/8) but all 5 samples from Burkina Faso mapped to lineage B.

Genomic diversity of HPV45 variants

Eleven variants from 217 HPV45-containing samples were selected for complete genome sequencing; they represented 4 women from Costa Rica, 3 from Rwanda, 2 from Zambia, and 2 from Burkina Faso. Classification of HPV45 variants was based on the analysis of these samples combined with 13 previously reported HPV45 genomes, including the reference isolate [18,49]. There were a total of 239 nucleotide sites amongst the 7858 bp HPV45 genome that were variable (3.0%). Of the 2440 encoded amino acids, 89 (3.6%) showed variations (Tables 1 and S2, Figure S1C). The maximum nucleotide difference was 1.5%; the most divergent isolates were Qv27565 (an A2 variant) and Qv35960 (a B1 variant); this distance represented the maximum diversity between lineage A and B variants (Table S2 and Figure 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. HPV45 variant tree topology and pairwise comparisons of individual complete genomes.

The phylogenetic tree was constructed as described in Figure 2. Distinct variant lineages and sublineages were determined as described in Figure 2. The percent nucleotide sequence differences are shown in the panel to the right of the phylogenetic tree as described in Figure 2.

https://doi.org/10.1371/journal.pone.0072565.g004

Isolate BF134 (sublineage A3) from a Burkina Faso woman had unique variations and was 0.57%-0.80% and 1.19% -1.21% different than the characterized A1/A2 and B1/B2 variants, respectively (Table S3). Based on the phylogenetic tree topology, BF134 separated from the precursor of the A1/A2 sublineages (Figure 4). Lineage A and B variants are distinguished by 42 SNPs across the genome (Figure S1C). In addition, several sets of SNPs are diagnostic for sublineages A1 (n = 3), A2 (n = 20), A3 (n = 20), B1 (n = 17), and B2 (n = 17) using the current 24 genomes (Figure S2).

HPV45 isolates from Costa Rica mapped to the A2 (37%, 64/171) or B1 (40%, 68/171) sublineages, samples from Thailand sorted to the B1 (62%, 5/8) or B2 (38%, 3/8) sublineages, samples from Rwanda were predominantly B2 (80%, 8/10), whereas samples from Zambia were predominantly A1 (45%, 10/22) and B2 (36%, 8/22) and those from Burkina Faso A1 (67%, 4/6).

Genomic diversity of HPV59 variants

Forty-five HPV59 isolates had the URR/E6 regions sequenced; six samples encompassing each unique variation pattern were selected for complete genome sequencing (Table 1). In total, 8 complete genomes, including the prototype [50] and one isolate from a Chinese woman [40] were analyzed. There were 169/7898 (2.1%) variable nucleotide positions and 66/2441 (2.7%) variable amino acids (Table S2 and Figure S1D). The maximum pairwise nucleotide difference was 1.3%, which was observed between isolates LZod68 and Qv33993 (Table S2); these divergent isolates were part of HPV59 sublineage A1 and lineage B, respectively, as inferred from the phylogenetic analysis (Figure 5). Isolates within lineage A were more variable than those from lineage B and were further classified into sublineages A1, A2 and A3, with mean differences of 0.60%-0.89% (Table S3). HPV59 lineages A and B had mean differences of 1.11%-1.26% and can be distinguished by 53 lineage-specific SNPs across the genome (Figure S2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. HPV59 variant tree topology and pairwise comparisons of individual complete genomes.

The phylogenetic tree was constructed as described in Figure 2. Distinct variant lineages and sublineages were determined as described in Figure 2. The percent nucleotide sequence differences are shown in the panel to the right of the phylogenetic tree as described in Figure 2.

https://doi.org/10.1371/journal.pone.0072565.g005

Inspection of the distribution of the forty-five HPV59 variants from 4 geographic regions indicated that 61% (20/33) of variants from Costa Rica were A2, whereas 67% (4/6) and 60% (3/5) of samples from Rwanda and Burkina Faso were of the B lineage, respectively (data not shown). Thus, the B lineage was more common in samples from Africa, but the numbers were limited and the difference was not statistically significant (p = 0.11).

Genomic diversity of HPV68 variants

Historically two isolates of HPV68 (HPV68Ref(R)/HPV68a and ME180/HPV68b) were characterized, shared 93% nucleotide sequence similarity within the L1 ORF, and were designated as subtypes [51,52]. We sequenced the URR and/or E6 regions of eighty-eight HPV68 positive samples from patients in Costa Rica, Taiwan, Rwanda and Burkina Faso, and identified 20 isolates (22.7%) related to HPV68R, and 68 (77.3%) related to ME180. All isolates (5 HPV68R-related and 12 ME180-related) representing unique variation patterns were sequenced. Additionally, four HPV68 complete genomes from NCBI/GenBank, including the 2 prototypes [51,52], are included in this report (Figure 6, Table 1 and Table S1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. HPV68 variant tree topology and pairwise comparisons of individual complete genomes.

The phylogenetic tree was constructed as described in Figure 2. Distinct variant lineages and sublineages were determined as described in Figure 2. The percent nucleotide sequence differences are shown in the panel to the right of the phylogenetic tree as described in Figure 2.

https://doi.org/10.1371/journal.pone.0072565.g006

The maximum pairwise difference, 6.7% across the complete genome and 7.8% within the L1 ORF, was between isolates from lineage B, Qv18016 (an HV68a variant) and lineage E, TJ42 (a ME180 variant) (Table S2). Overall, 9.8% (771/7850) of nucleotide sites and 9.6% (232/2419) of encoded amino acids were variable (Table S2, Figure S1E). Based on lineage assignment encompassing 1.0%-10.0% differences in complete genomes [12], the “subtype” designation was replaced by assigning variant lineages to this taxon. Hence, ME180-related genomes were assigned to HPV68 variant lineages C, D, E and F, and HPV68R-related isolates to lineages A and B.

HPV68 lineages A/B are distantly split from lineages C/D/E/F (Figure 6); all variants sorted into the two distinct groups and displayed mean differences of 6.38%-6.61% across the complete genomes (Table S2, Figure S2). Comparison of the intragenomic diversity of variants within clades A/B and C/D/E/F showed differences of 0.93%-0.97% and 1.20% to 1.72%, respectively (Table S2). Thus, the ME180-related variants were more variable than the HPV68R-related isolates, consistent with the phylogenetic analysis that two HPV68R-related lineages and four ME180-related lineages constitute HPV68 (Figure 6). The following sublineages further classify HPV68 isolates: A1 and A2, C1 and C2, D1 and D2, and F1 and F2 (Figure 6). The isolates from Burkina Faso mapped to lineages D2 (1/6), E (2/6) and F (3/6); the isolates from Rwanda mapped to A1 (2/14), B (1/14), C1 (7/14), E (1/14) and F (3/14); and all four isolates from Taiwan sorted to C2. Thus, 85% of isolates available from Africa were from the C/D/E/F clade, whereas 70% (47/67) of the isolates from Costa Rica were from the C/D/E/F clade (p = 0.19).

Genomic diversity of HPV70 variants

Amplification and sequencing the URR and/or E6 regions of 295 samples containing HPV70 clustered isolates into two distinct groups from which 8 representative isolates were selected for complete genome analyses, in addition to the reference sequence [52] (Table 1). In total, 192 nucleotide variable positions were observed across the 7922 bp genome (2.4%) and 67 of 2413 (2.8%) encoded amino acids were variable (Table S2). The maximum pairwise nucleotide difference was 1.6% between Qv27211 and Qv17574 (Table S2). Within the HPV70 genomes, we detected indels within the NCR2 and URR regions, and a 6-bp insertion (ACTGTA) between nt 6123 and 6124 within the L1 ORF resulting in an insertion of threonine and valine, between L1 aa position 178 and 179 (Figure S1F).

Phylogenetic trees inferred from HPV70 isolate nucleotide sequences separated variants into two distinct lineages, A and B (Figure 7). These lineages differ by 1.52 ± 0.13% and are distinguished by 82 lineage specific SNPs (Table S2, Figure S2). Isolates from the A lineage differed between geographic region: Rwanda 82% (18/22), Costa Rica 65% (175/268) and Burkina Faso 40% (2/5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. HPV70 variant tree topology and pairwise comparisons of individual complete genomes.

The phylogenetic tree was constructed as described in Figure 2. Distinct variant lineages and sublineages were determined as described in Figure 2. The percent nucleotide sequence differences are shown in the panel to the right of the phylogenetic tree as described in Figure 2.

https://doi.org/10.1371/journal.pone.0072565.g007

Alpha-7 HPV evolution

To investigate the relationship between isolates from the alpha-7 species group, we took one representative genome from each lineage or sublineage, constructed a phylogeny and plotted the percent genome differences from global alignments (Figure 8). Four clades were apparent from this analysis- HPV18/45/97, HPV59, HPV85 and HPV39/68/70 corresponding to the 4 deep-seated branches leading to these clades. Each clade was approximately 30% different from isolates within other clades, whereas the clades with more than one type (i.e., HPV18/45/97 and HPV39/68/70) consistently shared at least 80% of nucleotides. The intratype differences are displayed as a heatmap (Figure S3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 8. Phylogenetic tree and pairwise comparisons of representative Alpha-7 lineages and sublineage complete genomes.

The phylogenetic tree was constructed as described in Figure 2. Representative genomes for each variant lineage and sublineage described in this report are used and the names of the isolates are shown to the right of the tree branches. The percent nucleotide sequence differences are shown in the panel to the right of the phylogenetic tree as described in Figure 2.

https://doi.org/10.1371/journal.pone.0072565.g008

Alpha-7 HPV genomic diversity

Rarefaction curves of parsim-informative SNPs (sites detected in ≥ 2 samples) for combined genomes of each type were plotted to estimate the coverage of type specific variations within the sampled cohorts and genomes analyzed (Figure S4A). Although sampling of genomes within the targeted populations may increase the repertoire of genomic variability, it is unlikely to reveal novel variant lineages since the curves flatten out with increasing numbers of genomes already sequenced. Interestingly, the detection of singleton SNPs (variations present only once in the sampled genomes) increases almost linearly with increasing sample size of each type (Figure S4B). Rare variations add information on mutations that potentially occurred during recent epochs of explosive population growth or natural selection and may be identified from sequencing larger sample sizes [53]. As sequencing larger numbers of isolates becomes more economically feasible with Next-Gen technologies, false positive SNPs within the genomes of additional variants remains unavoidable due to errors introduced by PCR amplification, sequencing technologies, software analyses and other unknown causes of DNA sequence error.

The pairwise inter-lineage mean differences inferred from the genome sequences revealed maximum genomic diversity of HPV68 and HPV18, followed by HPV70, HPV45, HPV59 and HPV39. When each ORF/region was compared, the noncoding regions (NCR1, NCR2 and URR) most often showed the largest variability, followed by the E5 and E4/E2 overlapping ORFs (Table S2). The diversity of the E6, E7, E1, E2, L2 and L1 ORFs varied by type; nevertheless, E1 and L1 ORFs were more conserved in terms of the overall nucleotide diversity compared to the other ORFs (Table S2). Indel events were observed within both coding and non-coding regions, resulting in variability in length of specific ORFs/regions among variants of the same types (Figure S1 and Figure S5).