Whole Genome In-Silico Analysis of South African G1P[8] Rotavirus Strains before and after Vaccine Introduction over a Period of 14 Years

Rotavirus G1P[8] strains account for more than half of the group A rotavirus (RVA) infections in children under five years of age, globally. A total of 103 stool samples previously characterized as G1P[8] and collected seven years before and seven years after introducing the Rotarix® vaccine in South Africa were processed for whole-genome sequencing. All the strains analyzed had a Wa-like constellation (G1-P[8]-I1-R1-C1-M1-A1-N1-T1-E1-H1). South African pre- and post-vaccine G1 strains were clustered in G1 lineage-I and II while the majority (84.2%) of the P[8] strains were grouped in P[8] lineage-III. Several amino acid sites across ten gene segments with the exception of VP7 were under positive selective pressure. Except for the N147D substitution in the antigenic site of eight post-vaccine G1 strains when compared to both Rotarix® and pre-vaccine strains, most of the amino acid substitutions in the antigenic regions of post-vaccine G1P[8] strains were already present during the pre-vaccine period. Therefore, Rotarix® did not appear to have an impact on the amino acid differences in the antigenic regions of South African post-vaccine G1P[8] strains. However, continued whole-genome surveillance of RVA strains to decipher genetic changes in the post-vaccine period remains imperative.


Complementary DNA (Cdna) Synthesis
cDNA was synthesized from the extracted viral RNA using the Maxima H Minus Double-Stranded Synthesis Kit and protocol (Thermo Fischer Scientific, Waltham, MA, USA) with some modifications (UFS-NGS Unit SOP). Briefly, for first-strand cDNA synthesis, a denaturing step of the purified extracted RNA at 95 • C for 5 min in Multigene Optimax thermocycler (Labnet, Edison, NJ, USA) was followed by the addition of 1 µL of Random Hexamer primer, 100 µM. The reaction mixture was incubated in a thermocycler at 65 • C for 5 min. Afterward, 5 µL volume of the 4× First-Strand Reaction Mix and 1 µL of First Strand Enzyme Mix was added and the reaction mixture was incubated in a thermocycler pre-programmed as follows: 10 min at 25 • C, 120 min at 50 • C, and 5 min at 85 • C. For the second-strand cDNA synthesis step, a 55 µL volume of Nuclease-Free water, 20 µL of 5× Second Strand Reaction Mix, and 5 µL of Second Strand Enzyme Mix was added and incubated in the thermal cycler at 16 • C for 60 min. A 6 µL volume of EDTA, pH 8.9 was added followed by 10 µL of RNase I. The synthesized cDNA was purified using MSB ® Spin PCRapace Kit by following the manufacturer's protocol (Stratec Molecular, Berlin, Germany).

DNA Library Preparation and Whole-Genome Sequencing
The DNA libraries were prepared by utilizing the Nextera ® XT DNA Library Preparation Kit (Illumina, San Diego, CA, USA). Quantitative and qualitative assessment of DNA was performed using Qubit 3.0 fluorometer (Invitrogen, Carlsbad, CA, USA) and Agilent 2100 BioAnalyzer ® (Agilent Technologies, Waldbronn, Germany), respectively, by following the manufacturer's instructions. The DNA library and the PhiX Control v3 library (Illumina, San Diego, CA, USA) were normalized to 8 pM and 20 pM concentrations, respectively. A volume of 600 µL pooled denatured DNA library spiked with 20% PhiX Control v3 library was loaded into a MiSeq ® Reagent Kit V3 for paired-end nucleotide sequencing (301 × 2) on a MiSeq ® sequencer (Illumina, San Diego, CA, USA) at the University of the Free State-Next Generation Sequencing (UFS-NGS) Unit, Bloemfontein, South Africa.

Genome Assembly
The trimming of Illumina read ends and subsequent genome assembly was performed using a suite of tools embedded in Geneious Prime ® software, version 2020.1.1 [30]. Complementary RVA assembly was also performed using an in-house developed genome assembly pipeline and CLC Genomics Workbench 12 (https://www.qiagenbioinformatics.com/).

Generation of Whole-Genome Constellations
The genotype of each gene segment was determined using the RVA determination tool in the Virus Pathogen Database and Analysis Resource (ViPR) [31] to generate the full genome constellations for each RVA strain.

Phylogenetic Analyses
Alignments and comparative analysis of the full-length sequences for each gene segment was performed as described previously [24]. Multiple sequence alignments were performed using the MAFFT package in Geneious Prime 2020 [30]. Duplicated sequences in the alignments were identified utilizing ElimDupes (https://www.hiv.lanl.gov/content/sequence/elimdupesv2/elimdupes. html). The best evolutionary models for each gene segment were estimated using the DNA Model Test program in MEGA 6 to guide in the construction of maximum-likelihood trees with 1000 bootstrap replicates [32].

Selection Pressure and Recombination Analysis
Analysis of natural selection in RVA genome segments was done using the suite from the DataMonkey Webserver [33]: Fixed-effects Likelihood (FEL) [34], Fast Unconstrained Bayesian Approximation for Inferring Selection (FUBAR) [35] and mixed-effects model of episodic selection (MEME) [36]. Amino acid sites undergoing positive selection were identified and tabulated. Analysis of genetic recombination was performed using Genetic Algorithm for Recombination Detection (GARD) [34].

Protein Modeling
The RVA protein structures were modeled using SWISS-MODEL with an initial template search [37]. The templates were selected from the SWISS-MODEL Template Library (SMTL) and their respective resolution values for the analyzed genes were as follows: VP7 (3 fmg.1, 3.40 Å) and VP4 (2 dwr.1, 2.50 Å). The evaluation of stereochemical quality parameters of the generated structures was performed using the Structure Assessment feature in the SWISS-MODEL server [37] and VERIFY3D [38]. Image visualization and analysis was performed in PyMol software [39].

In Silico Analysis of Effect of Mutation(s) on Protein Stability
The FoldX plugin [40] integrated in the YASARA platform [41] was used to predict the stability effect of mutation(s) in a 3D structure. FoldX estimates the stability effect of a mutation empirically whereby the stability (∆G) of a protein is defined by the free energy, which is expressed in kcal/mol. In this study, G is the difference of free energy between the Rotarix ® (vaccine) strain or pre-vaccine strain and mutant strain. The energy with positive value is regarded to destabilize the structure, while a mutation with a negative value is regarded to stabilize the structure. Free energy change of ±0.5 kcal/moL is regarded as statistically significant for the stabilizing/destabilizing effect [40].

Whole-Genome Constellation Determination
The genotype constellation for the 103 South African G1P [8] strains (92 pre-and 11 post-vaccine) sequenced in this study and corresponding 68 strains (56 pre-and 12 post-vaccine) acquired from the GenBank database as reference strains was typical G1-P[8]-I1-R1-C1-M1-A1-N1-T1-E1-H1 (Supplementary Materials 1 (S1)). The sizes of the full-length genome segments one to eleven and their respective open reading frames (ORFs) were determined (S1). All the gene sequences in this study were submitted in the NCBI GenBank database under accession numbers MT854335-MT855467.

Phylogenetic Analyses of VP7 and VP4
Phylogenetic trees for each of the 11 gene segments were constructed. For VP7 and VP4, well-known lineage designations were utilized [9,11]. The VP7 phylogenetic tree comprised the VP7 gene sequences of the South African strains together with VP7 gene sequences of the reference strains from the seven established VP7 G1 genotype lineages [9]. The 171 G1 South African sequences utilized in this study were segregated into two main lineages: G1-lineage I and lineage II (Figure 1). G1-lineage I comprised 76 pre-vaccine G1 strains and 12 post-vaccine G1 strains, while G1-lineage II comprised 72 pre-vaccine G1 strains and 11 post-vaccine G1strains (Figure 1). Strain RVA/ Human-wt/ZAF/UFS-NGS-MRC-DPRU2250/2013/G1P [8] had ≥99.9% nucleotide identity with Rotarix ® in all its structural and nonstructural genes (S2) and clustered alongside cognate Rotarix ® genes in all the 11 phylogenetic trees. South African pre-and post-vaccine G1 strains were highly identical to each other and similarly distant from the Rotarix ® strain ( Table 1).
The VP4 phylogenetic tree comprised nucleotide sequences of the South African RVA strains and those of the reference strains from the four established VP4 P [8] genotype lineages (I to IV) [11] ( Figure 2). The P [8] strains for South African rotaviruses segregated into three evolutionary lineages, P[8]-lineage I, III, and IV ( Figure 2). Lineage III comprised a mixture of 144 pre-vaccine and 21 post-vaccine P [8] strains, while lineage IV comprised five pre-vaccine strains ( Figure 2). The P [8] strains identified pre-and post-vaccination were highly identical to each other and similarly distant from the Rotarix ® strain (Table 1).  South African pre-and post-vaccine G1P [8] strains were compared with each other and also compared with the Rotarix ® strain. The nucleotide identity values were calculated using the pdistance algorithm in MEGA 6 software (S2).  South African pre-and post-vaccine G1P [8] strains were compared with each other and also compared with the Rotarix ® strain. The nucleotide identity values were calculated using the p-distance algorithm in MEGA 6 software (S2). The VP4 phylogenetic tree comprised nucleotide sequences of the South African RVA strains and those of the reference strains from the four established VP4 P [8] genotype lineages (I to IV) [11] ( Figure 2). The P [8] strains for South African rotaviruses segregated into three evolutionary lineages, P[8]-lineage I, III, and IV ( Figure 2). Lineage III comprised a mixture of 144 pre-vaccine and 21 postvaccine P [8] strains, while lineage IV comprised five pre-vaccine strains ( Figure 2). The P [8] strains identified pre-and post-vaccination were highly identical to each other and similarly distant from the Rotarix ® strain (Table 1).

P[8]-Lineage-IV
Outgroup RVA/Human-tc/USA/DS-1/1976/G2P [4]   South African VP1-VP3, VP6, and NSP1-NSP5 were clustered into two main lineages, one of the lineages comprising of the Rotarix ® genome segments (Figures S1-S9 in S4). South African strains identified pre-and post-vaccination were highly identical to each other as well as similarly distant from the vaccine strain (Table 1). The VP7 protein contains two defined neutralization antigenic epitopes spanning across 29 amino acid residues: 7-1 (7-1a and 7-1b) and 7-2 regions [42]. A comparison of the amino acid residues in the antigenic epitopes of wild type South African G1 strains with cognate sites in the Rotarix ® strain was performed ( Figure S1 in S5). Five amino acid differences appeared in both pre-and post-vaccine strains, four amino acids occurred before vaccine introduction while two amino acid differences, N147D (in eight strains) and T242A (in one strain), appeared only during the post-vaccine period (Figure S1 in S5; Table S1 in S6). We performed protein modeling analysis on the post-vaccine strains, which showed amino acid differences relative to both the pre-vaccine strains and Rotarix ® strain. Therefore, for VP7, we selected strain RVA/Human-wt/ZAF/UFS-NGS-MRC-DPRU4357/2015/G1P [8] as a representative strain of the eight G1 post-vaccine strains that had N147D amino acid substitution ( Figure S1 in S5). Additionally, we selected strain RVA/Humanwt/ZAF/MRC-DPRU1544/2010/G1P [8], the only post-vaccine strain exhibiting a T242A amino acid difference ( Figure S1 in S5). The VP7 protein structures of the selected G1 strains were superposed with the Rotarix ® VP7 structure to assess differences in the structural conformation. The VP7 structure of strain RVA/Human-wt/ZAF/UFS-NGS-MRC-DPRU4357/2015/G1P [8] when superposed with the Rotarix ® VP7 had a root mean square deviation (RMSD) value of 0.020 Å (Table 2; Figure 3A). The RMSD value closer to zero suggests high structural homology [43]. Replacement of asparagine with aspartate (N147D) significantly destabilized the protein structure as the free energy change was +0.527 kcal/mol [40] (Table 2; Figure 3A). Apart from the VP7 protein structure of Rotarix ® , we performed protein modeling analysis using the protein structures of five randomly selected pre-vaccine study strains (Figure S1 in S7). Similar significant free energy change trends for the destabilizing effect were observed and ranged from +0.506 to +0.579 kcal/mol ( Figure S1 in S7). The VP7 structure of strain RVA/Human-wt/ZAF/MRC-DPRU1544/2010/G1P [8] had an RMSD value of 0.012 Å when superposed with Rotarix ® VP7 (Table 2; Figure 3B). Threonine, which is a neutral hydrophilic amino acid, was replaced with alanine, an aliphatic hydrophobic amino acid, showing a change in polarity. However, the replacing amino acid (alanine) does not significantly alter the stability of the protein at this epitope (−0.076 kcal/mol) [40]. Protein modeling analysis with VP7 protein structures from five randomly selected pre-vaccine study strains showed similar free energy change trends ranging from −0.070 to −0.078 kcal/mol ( Figure S2 in S7).   The structure of the study strain is in green, while the Rotarix ® structure is in deep-teal. The amino acid residues are represented in dark gray and red to examine whether the replacing amino acid alters the conformation of the protein structure. The amino acid residue highlighted in dark gray represents the study strain while that in firebrick red represents the Rotarix ® strain. RMSD is the superposition value where the value of zero indicates absolute similarity. The stability of the protein after mutation was measured in kcal/mol, whereby free energy change of ±0.5 kcal/moL was regarded as significant for the stabilizing/destabilizing effect. , respectively, that were only present in post-vaccine G1P [8] strains when superposed with VP7 of Rotarix ® . The structure of the study strain is in green, while the Rotarix ® structure is in deep-teal. The amino acid residues are represented in dark gray and red to examine whether the replacing amino acid alters the conformation of the protein structure. The amino acid residue highlighted in dark gray represents the study strain while that in firebrick red represents the Rotarix ® strain. RMSD is the superposition value where the value of zero indicates absolute similarity. The stability of the protein after mutation was measured in kcal/mol, whereby free energy change of ±0.5 kcal/moL was regarded as significant for the stabilizing/destabilizing effect.

Comparative Analysis of VP7 Cytotoxic T Lymphocyte Epitopes of South Africa With Rotarix ® Vaccine Strain
The two established VP7 cytotoxic T-cell lymphocyte epitopes at positions 16-28 and 40-52 [44] of the VP7 of South African G1P [8] strains were compared with cognate regions of Rotarix ® VP7. Five amino acid differences appeared before Rotarix ® introduction, while two amino acid substitutions (A43V in two strains and A46T in four strains) were found after vaccine introduction (Table S1 in S6). Comparative analysis of the South African post-vaccine G1 genotypes outside the known epitope regions with cognate regions in G1 of Rotarix ® identified three amino acid differences (S37I, A68D, and E222K) (S8). The replacing amino acid at positions S37I, A68D, and E222K resulted in a change in the polarity of the amino acid residue (Table S2 in S6).

Comparative Analysis of Neutralizing Antigenic Epitopes in VP4 Genes of South African P[8] Strains and Rotarix ® Vaccine Strain
The VP4 protein comprises the VP5* and VP8* regions. The VP8* region contains four (8-1 to 8-4) neutralizing antigenic epitopes, while VP5* has five (5-1 to 5-5) [45]. The differences between the P[8] study strains and Rotarix ® P[8] component were mostly contained in the VP8* epitopes 8-1, 8-3, and 8-4 ( Figure S2 in S5). Seven amino acid differences were identified in both pre-and post-vaccine strains, seven amino acid substitutions during pre-vaccination introduction and two amino acid substitutions, T88I (in one strain) and N89S (in one strain) were identified after vaccine introduction ( Figure S2 in S5; Table S3 in S6). We performed protein modeling analysis on the post-vaccine strains that showed amino acid differences relative to both the pre-vaccine strains and Rotarix ® strain and these were RVA/Human-wt/ZAF/UFS-NGS-MRC-DPRU74/2014/G1P [8] and RVA/Human-wt/ZAF/UFS-NGS-MRC-DPRU83/2011/G1P [8] (Figure S2 in S5). The VP4 structure of strain RVA/Human-wt/ZAF/UFS-NGS-MRC-DPRU74/2014/G1P [8] aligned with the RMSD value of RMSD 0.048 when superposed with the Rotarix ® VP4 (Table 3; Figure 4A). Replacement with the isoleucine at position T88I altered the polarity of the residue from polar to nonpolar without affecting the charge [46]. This change in polarity may alter the physical properties of the protein. The mutation did not significantly impact the stability of the protein (−0.297 kcal/mol) ( Table 3; Figure 4A). Further free energy change analysis with VP4 structures from five randomly selected pre-vaccine strains demonstrated similar free energy change trends that ranged from −0.504 to −0.322 kcal/mol ( Figure S3 in S7).The VP4 structure of strain RVA/Human-wt/ZAF/MRC-DPRU83/2011/G1P [8] had an RMSD value of 0.049 when superposed with Rotarix ® VP4, indicating significant alignment. Asparagine is a polar, neutrally charged amino acid, and the replacing amino acid, serine, is also polar with a neutral charge [46]. The replacing amino acid, serine, significantly destabilized the protein structure with +1.166 kcal/mol energy change (Table 3: Figure 4B).When protein modeling analysis was performed with VP4 structures of five randomly selected pre-vaccine P [8] strains, significant free energy change values ranging from +1.128 to +1.352 kcal/mol were also observed, indicating the amino acid substitution had a destabilizing effect on the protein structure ( Figure S4 in S7).    N89S (B), respectively, that were only seen in some post-vaccine strains when superposed with VP4 of Rotarix ® . The structure of the study strain is in green while the Rotarix ® structure is in deep-teal. The amino acid residues are represented in dark gray and red to examine whether the replacing amino acid alters the conformation of the protein structure. The amino acid residue highlighted in dark gray represents the study strain while the firebrick red represents the Rotarix ® strain. The stability of the protein after mutation was measured in kcal/mol, whereby the folding energy change of ±0.5 kcal/mol is regarded as statistically significant for either stabilizing (-)/destabilizing (+) effect.  N89S (B), respectively, that were only seen in some post-vaccine strains when superposed with VP4 of Rotarix ® . The structure of the study strain is in green while the Rotarix ® structure is in deep-teal. The amino acid residues are represented in dark gray and red to examine whether the replacing amino acid alters the conformation of the protein structure. The amino acid residue highlighted in dark gray represents the study strain while the firebrick red represents the Rotarix ® strain. The stability of the protein after mutation was measured in kcal/mol, whereby the folding energy change of ±0.5 kcal/mol is regarded as statistically significant for either stabilizing (−)/destabilizing (+) effect.

Analysis of the VP4 and VP6 Non-Neutralizing Regions
Comparative analysis of the South African post-vaccine P [8] strain outside the known epitope regions with cognate regions in pre-vaccine strains identified five amino acid differences that were conservative in nature (Table S4 in S6). Analysis of the amino acid differences between the VP6 gene sequences of South African pre-and post-vaccine strains identified an amino acid difference Y353H in a 2017 post-vaccine strain, RVA/Human-wt/ZAF/MRC-DPRU15948/2017/G1P [8] (S8). The amino acid mutation Y353H occurred in a conserved region of the VP6 that has been associated with trimerization and single-shelled particle formation [47] (Table S5 in S6).

Analyses of Selection Pressure and Recombination
Selection pressure analysis by Mixed-Effects Model of Evolution (MEME), Fast, Unconstrained Bayesian AppRoximation for inferring selection (FUBAR), and Fixed-Effects Likelihood (FEL) showed that most of the codon sites in the 11 gene segments were undergoing purifying selection. With the exception of the VP7 gene segment, several amino acid sites in the rest of the gene segments were identified to be under positive selective pressure (Table 4) Amino acid sites 7 and 245 were identified to be under positive selective pressure by all three analysis tools used (Table 4). GARD found no significant evidence of recombination in all gene segments.
Amino acid sites that were identified to be under positive selection by FEL, MEME, and FUBAR. Amino acid sites in bold and underlined were identified by all the three methods. For MEME and FEL, statistical significance was assessed at p ≤ 0.1 while for FUBAR, it was assessed at posterior probability ≥0.9. The dash (-) sign indicates no site was identified. Additional analyses reports are provided in S9.

Discussion
The present study reports whole-genome analysis of G1P [8] RVA strains collected seven years before and seven years after the introduction of the Rotarix ® vaccine in South Africa. Whole-genome analysis of South African G1P [8] strains showed that they possessed a Wa-like genotype constellation. These findings are comparable to what has been observed in several countries [11,[48][49][50][51]. The prevalent association of G1P [8] genotypes with the Wa-like genetic background has been hypothesized to be due to epidemiological fitness [52].
Of the seven previously described G1 lineages [9], South African pre-and post-vaccine G1 strains in this study were segregated into G1-lineage I and G1-lineage II. Similar observations have been reported in several other studies [11,51,53], highlighting the likelihood of the epidemiological fitness of G1 strains in these two lineages, which has led to their prevalent circulation over time.
Amino acid differences in the neutralization epitopes of VP7 compared to the cognate region in Rotarix ® identified at positions N94S, S123N, K291R, and M217T appeared in both pre-and post-vaccine strains and have been previously reported in Belgium and Brazil [51,61]. Since these amino acid substitutions were present even before the introduction of the vaccine, they can be attributed to the natural RVA evolutionary processes. Amino acid differences in N147D occurred in eight post-vaccine G1 strains resulted in a change from neutral to negatively charged amino acid residue, while T24A occurred in one post-vaccine G1 strain, which resulted in a change from hydrophilic to hydrophobic [46]. The change in charge may affect the protein's chemical properties, while the shift in polarity suggests possible inaccessibility of the epitope as it becomes more hydrophobic [46]. Differences in VP4 amino acids between the study strains and Rotarix ® were observed in epitopes 8-1 and 8-3, similar to those already reported in Belgium, Brazil, and Tunisia [51,61,62]. The polarity changes (hydrophilic to hydrophobic) observed at position T88I only in a post-vaccine strain may affect antibody binding at this region of the 8-4 epitope as the resulting epitope becomes relatively inaccessible due to the hydrophobic effect [46]. The loss of glycosylation site at position N89S only observed in the post-vaccine period in one strain may alter the function of the protein at the 8-4 epitope [63].
To gain further insight into the impact of amino acid substitution occurring in the neutralization epitopes of VP7 and VP4 on the stability of the protein structures, we performed folding free energy change analysis. A protein's folding free energy change is an essential aspect of the protein's stability with a direct association with the protein's function [64]. The N147D substitution in VP7 and N89S substitution in VP8* resulted in a predicted destabilizing effect on the protein structure. Due to this destabilization effect on the protein structure, as indicated by free energy change analysis, we hypothesize that these amino acid substitutions observed during the post-vaccine era may not be favored in the long-term to enhance viral fitness. We did not observe consistently occurring amino acid substitutions in the 11 gene segments in both the antigenic and non-antigenic regions throughout the post-vaccine period when contrasted to the pre-vaccine period. However, there were yearly Selection pressure analysis demonstrated that most of the codon sites in the 11 genome segments of South African G1P [8] strains were undergoing purifying selection, probably to purge deleterious polymorphisms that arise due to the inherent error-prone nature of the RNA polymerase enzyme [1].
Amino acid sites 7 and 245 in VP3 were identified to be under positive selection by all three selection pressure analysis methods and site 245 fell within the guanine-N7-methyltransferase (N7-MTpase) domain suggested to catalyze methylation during capping of nascent rotavirus transcripts [64]. Methylation capping by VP3 of RVA serves as a critical strategy to evade host antiviral innate immune response [65].

Conclusions
The study shows that there were no consistently conserved amino acid substitutions occurring throughout during post-RVA vaccine period in the antigenic and non-antigenic regions of South African G1P [8] strains as most of the amino acid substitutions found in the post-vaccine period were already present during the pre-vaccine period. Therefore, Rotarix ® did not appear to have an impact on the genetic changes of South African G1P [8] post-vaccine strains. However, continued long-term whole-genome surveillance to monitor any consistently occurring genetic changes during the post-vaccine period, which may hint at vaccine selective pressure, remains essential. This study was limited mainly by the significant differences in the sample size between the pre-and post-vaccine period, although the significant decline in G1P [8] strains in South Africa during post-vaccination period has been previously reported [21].  Figure S1: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 1 encoding VP1 protein, Figure S2: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 2 encoding VP2 protein, Figure S3: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 3 encoding VP3 protein, Figure S4: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 6 encoding VP6 protein, Figure S5: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 6 encoding NSP1 protein, Figure S6: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 8 encoding NSP2 protein, Figure S7: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 7 encoding NSP3 protein, Figure S8: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 10 encoding NSP4 protein, Figure S9: Maximum likelihood phylogenetic tree based on the full length sequences of the genome segment 11 encoding NSP5 protein. Supplementary data 5 (S5): Neutralization epitope analysis in VP4 and VP7 containing Figure S1: Alignment of VP7 neutralizing and cytotoxic T-cell lymphocytic epitopes residues of Rotarix ® strain against wild type South African G1P [8] strains, Figure S2: Alignment of antigenic residues in VP4 between the P[8] component of Rotarix ® and wild type P [8] strains. Supplementary data 6 (S6): Additional tables containing Table S1: Amino acid differences in the VP7 neutralization and cytotoxic T-cell lymphocyte (CTL) epitope regions between Rotarix ® and South African G1 strains, Table S2: Amino acid differences in South African post-vaccine VP7 Cytotoxic T-cell Lymphocyte epitope and non-epitope regions, Table S3: Amino acid differences in the VP4 neutralization epitope regions between Rotarix ® and South African P [4] strains, Table S4: Amino acid differences in South African post-vaccine P [8] RVA strains, Table S5: Amino acid differences in South African post-vaccine I1 RVA strains, Table S6: Amino acid differences in South African post-vaccine R1 RVA strains, Table S7: Amino acid differences in South African post-vaccine C1 RVA strains, Table S8: Amino acid differences in South African post-vaccine M1 RVA strains, Table S9: Amino acid differences in South African post-vaccine A1 RVA strains, Table S10: Amino acid differences in South African post-vaccine N1 RVA strains, Table S11: Amino acid differences in South African post-vaccine T1 RVA strains, Table S12: Amino acid differences in South African post-vaccine E1 RVA strains, Table S13: Amino acid differences in South African post-vaccine H1 RVA strains. Supplementary data 7(S7): Additional protein modeling analysis containing Figure S1: The VP7 structure models showing the N147D amino acid difference, Figure S2: The VP7 structure models showing the T242A amino acid difference, Figure S3: The VP4 structure models showing the T88I amino acid difference, Figure S4