Molecular Characterization Based on Whole-Genome Sequencing of Streptococcus pneumoniae in Children Living in Southwest China During 2017-2019

Background Streptococcus pneumoniae is an important pathogen causing high morbidity and high mortality in children and undergoes frequent recombination for capsule switching to neutralize the 13-valent pneumococcal conjugate vaccine (PCV13). This study aimed to investigate the prevalence, and molecular characteristics including serotypes and antibiotic susceptibility of S. pneumoniae isolated from children living in Southwest China from 2017 to 2019 to facilitate the selection of effective vaccine formulations and appropriate antibiotic treatment regimens. Methods This study was conducted at West China Second University Hospital (Chengdu, Sichuan Province, China), Zunyi Medical University Third Affiliated Hospital/First People’s Hospital of Zunyi (Zunyi, Guizhou Province, China) and Chengdu Jinjiang District Maternal and Child Healthcare Hospital (Chengdu, Sichuan Province, China). Demographic and clinical characteristics of children infected with S. pneumoniae were collected and analysed. Next-generation sequencing and sequence analysis were used to determine the serotypes, sequence types, antibiotic resistance and potential protein vaccine target genes of the pneumococcal isolates. The coverage rate provided by PCV13 was estimated by calculating the percentage of the specific serotypes that were specifically the PCV13-included serotypes. Antimicrobial susceptibility was determined by the microdilution broth method. Results The most prevalent pneumococcal serotypes were 19F (25.8%), 19A (14.1%), 6B (12.5%), 6A (9.4%) and 14 (7.8%). The predominant STs were ST271 (23.3%), ST320 (15.5%) and ST90 (8.6%), dominated by the clonal complex Taiwan19F-14 (39.1%). The coverage rate of PCV13 was 77.3% in all the isolates, with relatively higher values in invasive isolates (86.4%). Over the decade, the rates of resistance to penicillin, amoxicillin and cefotaxime were 5.6%, 5.3% and 5.1%, respectively, with significantly higher values in invasive isolates (22.4%, 14.9% and 11.9%). Almost all the isolates were resistant to erythromycin (99.1%) and clindamycin (95.9%). All isolates carried virulence-related genes, including ply, psaA, piaA, piuA, phtE, nanA, pepO, danJ, pvaA, clpP, pcsB, stkP, potD, and strH. The carriage of virulence and resistance genes varied among serotypes and clades, with serotype 19F/ST271 showing higher resistance to antibiotics and being more likely to carry pilus genes and other virulence genes. Conclusion These data provide valuable information for the understanding of pneumococcal pathogenesis, antimicrobial resistance and the development of protein-based vaccines against pneumococcal infection.


INTRODUCTION
Streptococcus pneumoniae (S. pneumoniae, also known as pneumococcus) is a gram-positive, extracellular, opportunistic pathogen that colonizes the mucosal surfaces of the human upper respiratory tract, and up to 27-65% of children and <10% of adults carry this organism (Abdullahi et al., 2012;Yahiaoui et al., 2016). This carriage is the prerequisite for both transmission to other individuals and disease progression in the carrier, including invasive pneumococcal diseases (IPDs, such as meningitis, septicaemia, and pleurisy) and non-invasive pneumococcal diseases (non-IPDs, such as pneumonia, otitis media, and sinusitis) (Weiser et al., 2018). The World Health Organization (WHO) estimated that approximately 800,000 children die of pneumococcal diseases annually, and more than 90% of these deaths occur in developing countries (O'Brien et al., 2009;Johnson et al., 2010).
There are at least 98 serotypes of S. pneumoniae circulating worldwide, categorized according to the unique glycan components and linkages that constitute the capsular polysaccharide of each serotype (Yother, 2011), which are the vaccine antigen targets of 13-valent pneumococcal conjugate vaccine (PCV13, including serotypes: 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F and 23F). However, geographical differences in serotype distribution lead to reduced effectiveness of the vaccine when they are implemented in geographic areas where the serotypes are not covered by PCV13 (Balsells et al., 2017). In addition, S. pneumoniae is known to undergo frequent recombination (Vos and Didelot, 2009;Mostowy et al., 2017), and capsule switching was reported to occur more frequently following the introduction of pneumococcal vaccines (Chang et al., 2015). Therefore, since PCV13 was introduced in mainland China in 2017, changes in serotype prevalence have been monitored to appropriately choose effective vaccine formulations.
The virulence of S. pneumoniae is conferred by capsular polysaccharides and multiple virulence factors, which may vary among clades. Knowledge of the virulence profile of isolates is crucial to predict disease severity and outcome of infection and allows risk assessment during the early onset of the disease (Deurenberg et al., 2017). Moreover, the resistance status is an important reference for antimicrobial selection for anti-infection therapy, and antibiotic resistance varies in different regions, so local antimicrobial resistance surveillance is essential to obtain evidence for clinical practice . Traditionally, studies on virulence and antibiotic resistance usually involve the design of primers for particular genes for PCR assays, which may limit the capacity of molecular characterization analysis. Currently, whole-genome sequencing technology based on nextgeneration sequencing (NGS) is used to determine the DNA sequence of a complete bacterial genome in a single sequence run, and from these data, information on virulence and antibiotic resistance, as well as information on molecular serotypes and sequence types (STs), is obtained, which is useful for pathogenic surveillance (Deurenberg et al., 2017).
In 2017, the WHO included S. pneumoniae as one of 12 priority pathogens, and the continued high burden of pneumococcal disease and rising rates of resistance to antibiotics have renewed interest in disease prevention (Weiser et al., 2018). Moreover, the pandemic of COVID-19 may have impact on the co-infection mode of S. pneumoniae and the virus and affect the molecular characteristics of the prevalent strains. Therefore, the baseline levels of the pathogen S. pneumoniae before the COVID-19 outbreak will provide important data to support and provide a reference for the investigation of the current status of pneumococcal diseases and possible genetic changes in subsequent prevalent isolates.
In this study, whole-genome NGS technology-based strategies were employed to investigate the prevalence and molecular characteristics of S. pneumoniae strains isolated from paediatric patients in Southwest China during 2017-2019. The coverage rates for PCVs were calculated. The virulence protein encoding genes and antibiotic resistance-related genes among S. pneumoniae serotypes/STs were analysed and compared.

Study Area and Population
This study was conducted from March 2017 to November 2019 at West China Second University Hospital (Chengdu, Sichuan Province, China), Zunyi Medical University Third Affiliated Hospital/First People's Hospital of Zunyi (Zunyi, Guizhou Province, China) and Chengdu Jinjiang District Maternal and Child Healthcare Hospital (Chengdu, Sichuan Province, China). These hospitals are one of China's largest specialty hospitals for children and women, a general hospital affiliated with a provincial medical university, and a typical community healthcare hospital, respectively. The hospitals' clinical laboratories have been accredited by the College of American Pathologists (CAP) or the China National Accreditation Service for Conformity Assessment (CNAS) under the ISO15189 accreditation standard or are under the supervision of the above-mentioned external quality assessment laboratory.
The enrolled subjects were children from Southwest China presenting with an S. pneumoniae infection who were admitted to these hospitals. The participant eligibility criteria included the following: (1) was younger than 14 years old; (2) gave clinical specimens from which S. pneumoniae was isolated and positively cultured; (3) had respiratory, neural, circulatory or local infectious manifestations; and (4) was not vaccinated against S. pneumoniae.
To have a comprehensive understanding on the antibiotic susceptibility of S. pneumoniae, we expanded the sample size by retrospectively collecting additional antibiotic susceptibility test (AST) results of S. pneumoniae isolates from patients (retrospective cohort) that met the participant eligibility criteria between January 2010 and March 2017.

Isolation and Identification of Strains
The strains of S. pneumoniae were collected, isolated and identified in line with the requirements for clinical procedures as previously reported (Yan et al., 2019). In brief, specimens were collected by specialized sample collection personnel or physicians, and the strains were isolated on Columbia agar + 5% sheep blood plates (BD Medical Technology, NJ, USA), which were incubated at 35°C for 24-48 h in a 5% carbon dioxide (CO 2 ) environment. All isolates were identified by typical colony morphology and optochin assays, and the results were confirmed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS; Vitek MS system; BioMerieux, Rhone, France). Strains isolated from sterile sites e.g. cerebrospinal fluid, blood and pleural fluid, were defined as IPD strains. Pneumococcal isolates were stored in 25% sterile glycerol broth at −70°C for subsequent analysis.

Genome Sequencing, Assembly, and Annotation
Genomic DNA was extracted by the QIAamp DNA Minikit (Qiagen, Hilden, Germany), and sequencing libraries were generated using the NEBNext ® Ultra ™ DNA Library Prep Kit for Illumina (New England Biolabs, NEB, USA) following the manufacturer's recommendations. Then, the whole genomes of S. pneumoniae were sequenced using the Illumina NovaSeq PE150 platform (Illumina, San Diego, CA, USA) with approximately 200× coverage at Beijing Novogene Bioinformatics Technology Co., Ltd. The genome data were assembled with SPAdes software (v3.14.1) (Prjibelski et al., 2020), and different Kmers (21,33,55,77) were selected for assembly to obtain the assembly result with the optimal k-mer value and the fewest scaffolds. All of the assembled genomes were submitted to GenBank and approved (PRJNA656156, PRJNA681770 and PRJNA643306). The genome of S. pneumoniae R6 (GenBank: AE007317.1) was used to annotate the assembled genome with Prokka software (v1.14.5) (Seemann, 2014).

Statistical Analysis
Statistical Package for Social Science (SPSS) software for Windows was used to assess the statistical significance of the data (version 22.0; Chicago, IL, USA). The chi-square test, Fisher's exact test and T-test were used. On the basis of the chi-square test, the Bonferroni method was used to test whether the differences among multiple groups were statistically significant, and P values < 0.05 were considered statistically significant.

Ethics Statement
The clinical experimental plan was approved by the Clinical Trial Ethics Committee of West China Second University Hospital, Sichuan University (No. 2018021). Before enrolment, written informed consent was obtained from legal guardians on behalf of the children involved in the study. The work was carried out in accordance with the Declaration of Helsinki.

Demographic and Clinical Characteristics
A total of 128 patients (prospective cohort) from 26 cities were enrolled from March 2017 to November 2019. Fifty-nine IPD cases and 69 non-IPD cases were included, with ages ranging from 0 to 14 years and a median age (P25-P75) of 1.17 (0.75-3.00) years. The pneumococcal diseases included meningitis, bacteraemia, pleural and peritoneal inflammation, pneumonia, bronchopneumonia, upper respiratory tract infections and otitis media. In total, 121 patients (94.5%) had a favourable prognosis.
To have a comprehensive understanding on the antibiotic resistance phenotype, we retrospectively collected additional AST results of pneumococcal isolates from 993 cases (retrospective cohort) admitted at hospitals between January 2010 and March 2017. Overall, 926 non-IPD cases and 67 IPD cases were included, with a median age (P25-P75) of 1.00 (0.58-2.00) years ( Table 1). The demographic and clinical characteristics of the patients were summarized in Table 1. No significant differences in those characteristics were observed between the two cohorts.
The relationship between the isolates and virulence genes is shown in the annotation of the RAxML tree ( Figure 4). Gene carriage rate varied significantly among strains of different serotypes. For adherence factors, the carriage rate of cbpG in serotype 6B was significantly higher than that of other serotypes (93.8% vs. 27.7%, p=0.000), and the carriage rate of PI-1 was highest in serotype 19A/19F (88.2% vs. 3.9%, p=0.000), which was similar to that of PI-2 (94.1% vs. 2.6%, p=0.000). pfbA was rarely detected in serotype 19A/19F but was prevalent in other serotypes (5.9% vs. 97.4%, p=0.000). For exoenzymes, the carriage rate of nanB was lowest in serotype 6B but highest in serotype 19A/19F/14 (31.3% vs. 84.3% vs. 100.0%, p=0.000). For protease, the carriage rate of iga was highest in serotype 19F but lowest in serotype 14 (100.0% vs. 57.6% vs. 0.0%, p=0.000).     a The classification of disease types combines clinical manifestations and pathogenic diagnosis and ultimately depends on the source of the strain. All patients identified as IPD had S. pneumoniae isolated from their normally sterile sites (cerebrospinal fluid/blood/pleural fluid/ascites). b When rrgA, rrgB, rrgC, srtC-1, srtC-2 and srtC-3 were all detected, the isolate was judged as pilus-Ⅰ (PI-1) positive. When srtA, pitB, sipA, srtG1, and srtG2 were all detected, the isolate was judged as pilus-Ⅱ (PI-2) positive. % Indicates the proportion of the gene detected in the subgroup. Bold indicates that the difference is statistically significant. On the basis of the chi-square test, the Bonferroni method was used to test whether the differences among multiple groups were statistically significant, and the symbols †, § and are used to indicate groups with significant differences ( §, high; ¶, medium; †, low). * indicates P value < 0.05.
The carriage rate of virulence genes also varied according to CC. For adherence factors, the carriage rate of cbpG was highest in CC90 but lowest in CC271/CC876 (100.0% vs. 60.3% vs. 0.0%, p=0.000), and lytB in CC271/CC90/CC876 was significantly higher than that in other serotypes (100.0% vs. 82.8%, p=0.004). The carriage rate of PI-1 was highest in CC271 (94.0% vs. 1.3%, p=0.000), which was similar to that of PI-2 (100.0% vs. 0.0%, p=0.000). However, pfbA was rarely detected in CC271 but was prevalent in other CCs (0.0% vs. 100.0%, p=0.000). For exoenzymes, the carriage rate of nanB was lowest in serotype CC90 but highest in CC271/CC876 (0.0% vs. 86.2% vs. 100.0%, p=0.000). For protease, the carriage rate of iga was highest in CC271/CC90 but lowest in CC876 (96.7% vs. 39.7% vs. 0.0%, p=0.000). It was found that pilus and pfbA were mutually exclusive, while PI-I and PI-II were closely linked ( Figure 3C). All strains in this study carried either PI-II or pfbA but not both. In addition, the presence of virulence genes did not show a significant difference among the isolates obtained from IPD and non-IPD patients in this study.
Amino acid substitutions in conserved catalytic motifs (SXXK) of the penicillin-binding proteins (PBPs) were observed in all b-lactam non-sensitive strains, as well as some

DISCUSSION
In this study, we describe the application of whole-genome NGS approach for comprehensive identification of the molecular characteristics including serotype, ST, conservation of virulence factor gene and the presence of antibiotic resistance gene of   (Suaya et al., 2020), the UK and Ireland (3, 8, 15A) (Pick et al., 2020), and Japan (12F, 3, 23A) (Yanagihara et al., 2021). The similarity or discrepancy in the distribution of serotypes may relate to the following factors: 1) native S. pneumoniae circulating in various regions has different evolutionary directions, resulting in different capsule genes; 2) population of different races has different susceptibilities to different serotypes of S. pneumoniae; 3) the PCV vaccination status of the populations in different regions varies, which may be the most important factor. Serotype replacement has been observed in regions where PCVs are widely available (Richter et al., 2013;Nakano et al., 2016). Compared with our previous data (Yan et al., 2019), the

A B
FIGURE 5 | Isolate collection and antibiotic resistance of S. pneumoniae. (A) Proportion of cases from which S. pneumoniae was isolated in each month (n=993). JAN: 10.2%, FEB: 7.2%, MAR: 9.0%, APR: 11.2%, MAY: 7.8%, JUN: 6.9%, JUL: 6.2%, AUG: 5.3%, SEP: 3.5%, OCT: 8.0%, NOV: 11.9%, DEC: 12.9%.  (Wang et al., 2020a;Wang et al., 2020b) but remains higher than that in other developed countries (Spain: 52%; USA: 41.4%; Japan: 37.5%) (Richter et al., 2013;Nakano et al., 2016;Ciruela et al., 2018). In agreement with the results reported by Wang et al. (Wang et al., 2020a), the present study also shows that the serotype coverage for PCVs was higher for IPD strains than non-IPD strains, and serotype 14 strain was several times more frequently isolated in IPD patients than non-IPD patients. These results suggest that the PCV13 vaccine remains an effective means to control pneumococcal infection in this region, especially for invasive infection. The present study shows that ST271, ST320, ST90 and ST876 were the most prevalent STs in Southwest China, which were similar to those in other regions in China (Li et al., 2018;Wang et al., 2020b). The present study further revealed 8 novel STs which accounted for 6.25% of all isolates (8/128), suggesting a higher variation and mutation rate for S. pneumoniae over other respiratory pathogenic bacteria e.g. S. aureus and H. influenzae. Based on SLVs as a standard, the predominant prevalent PMEN clone was Taiwan 19F -14, which constituted more than one-third of the total and was twice as prevalent in this region as in Shanghai (18.7%, 14/75) . These results suggest that S. pneumoniae may have multiple origins in China. In addition, the prevalent STs in China were different from those in other countries, such as Japan (Nakano et al., 2016), northern Russia (Vorobieva S Jensen et al., 2020), Iran (Talebi et al., 2016) and Latin America (Moreno et al., 2020), strongly suggesting the diversity of S. pneumoniae on a global scale.
S. pneumoniae harbours dozens of virulence factors, including the capsular polysaccharide, and adhesion factors e.g. exoenzymes, IgA1 protease, metal ion uptake factors, protease, and toxin factors, which were implicated in the pathogenesis of pneumococcal disease. Colonization is thought to be the first step for host infection (Weiser et al., 2018), and some of the key factors include: 1) pilus-1 (RrgA) and PfbA, which are implicated in first contact with the epithelium and epithelial receptors; 2) Phts, Eno and PepO, which can interact with the complement system; 3) NanA, BgaA and StrH, which function in mucus degradation; 4) PiaA, PiuA and PsaA, which harbor metal binding activity; 5) CbpE, which is related with the impairment of neutrophil activity; and 6) pneumolysin (Ply), which can damage the epithelium cells and have pro-inflammatory effects. The present work demonstrate that all isolates carry the above key factor genes, namely, phtE, eno, pepO, nanA, bgaA, strH, piaA, piuA, psaA, cbpE (pec) and ply, and carried at least one rrgA and pfbA, suggesting their indispensability in the pathogenic process of the isolates. After colonization, a variety of other virulence factors are involved in the following processes of pathogenesis, including adherence-related proteins e.g. LytA/B/C, CbpD/G, Lmb, SrtA, SlrA, and NanB, immunomodulatory proteins e.g. Plr, CppA, IgA1 protease and ZmpB/C, invasiveness-related proteins e.g. PavA, Pilus-2, HysA, HtrA and Tig. In the samples of this study, the carriage of multiple virulence genes was associated with the serotypes and CCs of the isolates. It was found that the serotypes prevalent in this region exhibited high carriage rates for nanB (19F/19A/14),iga (19F), cbpG (6B) and pfbA (6B/6A/ 14), which may be the result of natural selection in regions with low vaccination rates. The CCs prevalent in this region showed high carriage rates for lytB (CC271/CC90/CC876), Pilus-1 (CC271), Pilus-2 (CC271), nanB (CC271/CC876), iga (CC271/CC90), cbpG (CC90) and pfbA (CC90/CC876), which may be the reason why these branches were common in clinical infections.
Some virulence factors of S. pneumoniae are also potential targets for recombinant protein vaccines. Compared with PCVs, recombinant protein vaccines have lower cost and stronger immunogenicity and are not restricted by serotypes (Masomian et al., 2020). According to the target gene sequences of recombinant protein vaccines reported in the literature, we examined 27 protein vaccine candidate genes and found that more than 98% of the strains carried these genes, and 16 genes had sequence identity more than 98%. These results suggest that these protein antigens are widely conserved in pneumococcal isolates in Southwest China and may have high application potential in the development of recombinant protein vaccines.
PEN has historically been the first-line choice of antibiotics to treat S. pneumoniae infections. The present work demonstrates that the resistance rate of the isolates to PEN was less than 5% in non-meningitis patients, coincident with the results reported in other regions in East Asia (China: 0.7%-2.2%, Korea: 1.0%, Japan: 1.1%-3.8%) (Toda et al., 2018;Wang et al., 2019;Kim et al., 2020). However, 50%-70% of the meningitis strains were resistant to PEN using the MIC breakpoint ≥ 8 mg/mL according to CLSI 2020 criteria. These results suggest that PEN can continue to be used as a therapeutic drug for non-meningitis S. pneumoniae at higher concentrations, but it is not suitable for meningitis strains.
Altered PBP1a, PBP2b and PBP2x are the most important PBPs for b-lactam resistance among clinical isolates (Hakenbeck et al., 2012;Zhou et al., 2016). In this study, amino acid substitutions in conserved catalytic motifs STMK of PBPs were found in all blactam non-sensitive strains. The consistency, sensitivity and specificity of using the PBP2bAla617, PBP1aT540 and PBP1aT371 substitutions to predict MIC values less than 0.06 mg/mL were 96.3%, 100% and 84.6%, respectively. These results suggest that detection of the mutations may be helpful to predict the PEN resistance phenotype of isolates. It was found that IPD strains were more likely to have AMX and cephalosporins resistant phenotype than non-IPD strains, and the resistant rate of IPD strains to commonly used cephalosporins has also increased over the decade. These results suggest that MEM in combination with VAN can be chosen for the treatment of IPD before antibiotic susceptibility results are issued.
Macrolides were once heavily used in China because they were empirically used to cover Mycoplasma pneumoniae in children with infectious diseases and because of their safety in children and the exemption of a skin test before medication. According to the data from the China Antimicrobial Resistance Surveillance System in 2019, the resistance rate of S. pneumoniae to ERY was very high, with an average of 95.6% across the country. Macrolide resistance in S. pneumoniae is mainly mediated by: 1) target modification by a ribosomal methylase encoded by the erm(B) gene; and 2) a membrane efflux pump encoded by the mef(A/E) gene. The erm(B) gene was commonly detected in populations in mainland China, Taiwan, Japan, Thailand, Sri Lanka and South Korea, while mef(A) was more commonly detected in populations in Hong Kong, Singapore, Philippines, India and Malaysia (Kim et al., 2012). In this study, erm(B) and mef(A) were detected in 98.4% and 41.4% of the strains, respectively, which may be responsible for the heavy ERY resistance burden in this region. Although 14 ministries of China jointly issued a 5-year national action plan to control antimicrobial resistance (NAP) in 2016, 51.4% of prescriptions related to antibiotics, including the use of macrolides, remain inappropriate (Zhao et al., 2021). Similar to the results from other reports Wang et al., 2020b), the isolates were highly resistant to CLI in this work, thus, CLI may be not suitable for the treatment of S. pneumoniae infections. In addition, due to the potential toxic effects on children, the use of CHL (bone marrow suppression), TET (enamel hypoplasia), quinolones (achondroplasia) and sulphonamides (jaundice) is prohibited or restricted at different ages. These data suggest that macrolides are no longer suitable for the treatment of S. pneumoniae infections in China.
In this study, correlations between STs (CCs) and serotypes, virulence genes and antimicrobial resistance-related genes were observed, and some STs showed obvious aggregation in the branches of the RAxML tree. ST271, ST320, ST90 and ST876 were specifically associated with serotypes 19F, 19A, 6B and 14, respectively, which was quite similar to the results in a previous report (Li et al., 2018). Serotype 19F/ST271, as the predominant prevalent strain in the region, showed higher PEN nonsensitivity and was more resistant to CTX, CRO, and SXT, which is consistent with Margaret's report (Ip et al., 2015). Moreover, serotype 19F/ST271 was more likely to carry the virulence factors encoding genes pilus-1, pilus-2, nanB and iga as well as the macrolide resistance genes mef(A) and msr(D). These results suggest that some strains or clones may have evolved many important features including resistance to antibiotics and pathogenesis-related strategies to have a competitive advantage in adapting to the host. Strategies e.g. vaccination are therefore needed to control the widespread of these prevalent strains or clones.
The main strengths of the present study are the multicentre design and the utility of whole-genome NGS technology to characterize the prevalence of pneumococcal strains and analysed the encoding genes associated with pneumococcal virulence factors, potential protein vaccine candidates, and the antibiotic resistance proteins.
This study has some limitations. First, the sample size is relatively small which may undermine the representativeness of the prevalent strains in Southwest China. Second, since the strains isolated during 2010-2017 were unavailable, the evolution of genes associated with drug resistance cannot be compared and analysed. Third, only a very limited data mining was currently performed, further studies regarding the correlations between the genome characteristics and strain phenotypes or clinical indicators are warranted.
In conclusion, serotypes 19F, 19A, 6B, 6A, and 14 were most prevalent strains responsible for childhood pneumococcal disease in Southwest China. Some strains or clones, e.g. Taiwan 19F -14 have evolved many antibiotic resistance strategies against PEN, CTX, CRO, and SXT. Given the high coverage rate of PCV13 and the worrisome non-susceptibility rate of pneumococcal isolates to antibiotics, vaccination with PCV13 may be beneficial in this region. Since pneumococcal virulence protein encoding genes are highly conserved in pneumococcal isolates, the development of virulence protein-based vaccines is also expected.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Clinical Trial Ethics Committee of West China Second University Hospital, Sichuan University. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. (No. 17ZD005,No. 19PJ227 The funders had no role in study design, data collection and analysis and interpretation of the data.