Next generation deep sequencing and vaccine design: today and tomorrow.

Next generation sequencing (NGS) technologies have redefined the modus operandi in both human and microbial genetics research, allowing the unprecedented generation of very large sequencing datasets on a short time scale and at affordable costs. Vaccine development research is rapidly taking full advantage of the advent of NGS. This review provides a concise summary of the current applications of NGS in relation to research seeking to develop vaccines for human infectious diseases, incorporating studies of both the pathogen and the host. We focus on rapidly mutating viral pathogens, which are major targets in current vaccine research. NGS is unraveling the complex dynamics of viral evolution and host responses against these viruses, thus contributing substantially to the likelihood of successful vaccine development.

Vaccine development

Immunization is one of the most effective and sustainable ways of preventing infectious diseases, and also some cancers associated with infection. Vaccines against diphtheria, tetanus, poliomyelitis, influenza, hepatitis B, measles, mumps, rubella, as well as pneumococcal, meningococcal, and Haemophilus influenzae B infections, have reduced the incidence and mortality of these infectious diseases by greater than 97–99% [1]. However, there are still many medically significant human infectious diseases for which no vaccine exists.

The key constraint in the development of vaccines for protection against many prevalent human pathogens has been the variability in pathogen genomes across epidemics, or even within a single host infection episode. In addition, there has been a lack of understanding of how these microorganisms evolve to escape host immune responses, or conversely insufficient insight into the characteristics of protective immunity exemplified by the failure of HIV vaccines [2]. At the forefront of this conundrum are infections caused by RNA viruses, which are the most common pathogens of humans and animals, and largely have no effective vaccines available. These viruses include the high-profile epidemic pathogens, such as HIV-1, Hepatitis C virus (HCV) and Dengue virus, as well as endemic viruses that particularly cause morbidity and mortality among infants and aged individuals, such as norovirus and enteroviruses. In addition, emergent RNA viral pathogens are a major concern, typified by swine flu (influenza A H1N1), which arose via genetic recombination and then crossed species barriers to become pandemic in 2009 [3].

Since 2005, the development of high throughput, or so-called NGS technologies, has allowed a massive increase in capacity to sequence genomes at a relatively low cost and in a short time frame. NGS refers to a collection of high-throughput sequencing technologies developed since 2005, which use amplification-based assays to sequence in parallel many genomes from individual templates [4] (Box 1 ). The current NGS technologies are known as second generation technologies (Box 1, Table I), to differentiate them from the first generation (Sanger sequencing), and the third generation – which is based on single molecule sequencing [5].

Table I

Representative NGS sequencing platforms and their characteristicsa

Platform	Run time (h)	Read length (bp)	Throughput per run (Mb)	Typical errors	Main biological applications	Company URL
Roche 454 FLX +	23 hours	700, up to 1000	700	Insertions/deletions (indels) at homopolymer regions	Microbial genome sequencing, human genome sequencing, transcriptomics, metagenomics	http://www.my454.com/
IlluminaHySeq 1000MySeq 2000 V3	810	2 × 1002 × 150	400,000<600,000	Indels, especially end of reads	Microbial genome sequencing, human genome sequencing, transcriptomics, metagenomics	http://www.illumina.com/systems.ilmn
SOLiD 4	12	50 × 35	71,000	End of read substitution errors	Microbial genome sequencing, human genome sequencing, transcriptomics, metagenomics	http://www.appliedbiosystems.com/absite/us/en/home.html
Ion torrentPGM 318 Chip	3	200	1000	Indels at homopolymer regions	Microbial genome sequencing, human genome sequencing, transcriptomics, metagenomics	http://www.iontorrent.com/
Pacific Biosciences				Random indel errors	Full-length transcriptomics, discovering large structural variants and haplotypes	http://www.pacificbiosciences.com/

Data taken from web sites of the NGS companies.

Current NGS technologies that are dominating the market are known as second generation sequencing technologies, to separate them from the first generation of sequencing assays based on the Sanger method. The techniques involved in NGS include template preparation, sequencing and imaging, and data analysis (see Box 3 and [4] for detailed review). The combination of specific protocols for these techniques distinguishes one available technology from another.

Template preparation involves randomly breaking the DNA of interest into small fragments, which are then attached to a support such as beads in suspension, or a solid interface. Spatially separated immobilization of these template sites allows thousands to billions of sequencing reactions to be performed simultaneously in a process consisting of successive incorporation, washing and scanning operations to capture the sequence data. Second generation techniques are typically based on obtaining sequences from clonally amplified templates, whereas third generation techniques use single molecules with multiple nucleotide or probe additions [5] – both options generate technical errors.

Commercially available NGS technologies differ in the coverage, read length and specific chemical technologies used to sequence and read the generated strands (Table I ). For instance, Roche 454 Titanium technology has moderate coverage (or capacity) (∼1Gb per run), but has an average read length of 450 nucleotides (nt) and a maximum of ∼800 nt. By contrast, Illumina or SoliD technologies have a much higher coverage (∼20–50 Gb per run), but an average read length of <15 nt. Most of the technologies also offer the possibility of paired reads – a mechanism to link two separate reads across the genome. This is improving the quality and applicability of NGS (Table II ).

Table II

Major current applications of NGS technologies

NGS method	Application	Vaccine relevance	Refs
Genome sequencing	Genomics	Detection of genetic variationMetagenomicsDiscovery of new pathogen genomesImmune escapeVaccine safetyDiversity of T and B cell repertoireGenotyping	12, 16[60][14][48][53]43, 61[19]
RNA sequencing (RNA-Seq]	Transcriptomics, abundance analyses, analysis of non-coding RNAs	Immune regulationHost–pathogen interactionsmiRNA	39, 62, 6364, 65, 66, 67
Chromatic immunoprecipitation and sequencing ChIP-Seq	Global profiling of the epigenome,DNA–protein binding network	Immune regulationEpigenetics	26, 27[68]

The most commonly used NGS technologies at present are: (i) amplification of DNA material via pyrosequencing (Roche 454); (ii) reversible dye-termination sequencing (Illumina); and (iii) sequencing by ligation (SOLiD). This field is rapidly expanding, and novel improved platforms are continuously being developed and released. Examples include Heliscope by Helicos (http://www.helicosbio.com/), Ion Torrent Life Technologies (http://www.iontorrent.com/) and a real-time sequencing platform by Pacific Biosciences (http://www.pacificbiosciences.com/). Third generation techniques, based on sequencing of a single molecule of DNA or RNA, without intermediate steps between reading two segments are rapidly emerging [5].

In this review we consider current NGS applications that have relevance for vaccine research, especially where a systems biology (see Glossary) approach is being undertaken.

Current applications of NGS technologies

The key elements of the NGS revolution are: (i) the depth at which mixed populations of DNA or RNA genomes are sequenced; (ii) the volume of data; and (iii) the high-throughput capability, which allows rapid and direct measurement of whole genomes or transcriptomes. Therefore, rather than focusing on individual, detailed phenomena (without a priori knowledge of their importance), NGS allows systems approaches. Today, the major applications of NGS (Box 1, Table II) are: genome sequencing, transcriptome analysis (RNA-Seq), DNA–protein binding analysis (ChIP-Seq), and histone modification (NGS-methylation) (for a comprehensive review of these technologies, see [4]). For vaccine research, NGS has numerous applications (Figure 1 ), including sequencing of host and pathogen genomes [6] and their transcriptomes 7, 8, as well as studies of the diversity in host immune responses, in both T and B cells 9, 10, 11. The analysis of NGS data requires complex computational and bioinformatics techniques, as these datasets carry important limiting factors; notably, limited read lengths and high technical error rates (Box 2 ).

Figure 1

Next generation sequencing (NGS) is applicable to a wide spectrum of settings with a direct impact on vaccine research. Applications of NGS for vaccine studies range from systematic analyses of many samples collected from human populations, to detailed longitudinal studies of host–pathogen interactions within fewer subjects. NGS allows rapid assessment of both human and pathogen genomes, their transcriptomes, as well as examination of host immune responses, such as T and B cell diversity. NGS can be used to assess the quality of vaccine stocks, the diversity of HLA polymorphisms in large populations, and also for detection of new pathogenic strains in mixed samples.

A key challenge in NGS technology is the bioinformatics and statistical analysis of the datasets generated to ensure high quality in the analyses and interpretation of large, error-prone data sets. The increasing size of the NGS datasets being generated, short read lengths, and significant technical error rates carried with each of the emerging technologies will continue to demand sophisticated and efficient support systems 4, 5.

Notably, application of NGS analysis to study genome diversity and its structure can be significantly hampered by the current methodology involved in the sample preparation. For instance, for RNA viruses, the process often involves reverse transcription and PCR amplification, both of which can introduce significant bias in terms of point mutation and recombination events in the output sequence. This is particularly relevant as RNA viruses have a highly diverse population and it is important to distinguish low frequency variants from technical errors. For example, it is possible that with an error rate of 0.2% per base copied, more than half of the reads of a 454 run may carry at least one error (assuming an average read length of 400 bp) [55].

Quantitative methods are necessary to achieve a complete understanding of NGS data. For instance, the analysis may be used to inform mathematical models that describe pathogen evolution both within host and between hosts [57]. There are still several limitations in the application of NGS, such as the relatively short read length (see Box 1, Table I), and a very high error rate, which compromise at least in part the quality and range of potential applications. For instance, the challenge of reconstructing sequence haplotypes kilobases in length from short read NGS data is substantive [55]. New bioinformatics tools are being developed to allow reassembly of such haplotypes from viral genomes, and early steps have been already taken for achieving this goal in diploid genomes [58]. The dedicated NGS bioinformatics research field is growing rapidly, thereby providing the research community with advanced algorithms and accessible tools for more accurate data analyses (Box 3).

NGS and genomics

The first breakthrough in the application of NGS to vaccinology was sequencing the full genome of organisms and their hosts within hours to days at moderate costs. Since 2008, when the first whole human genome sequence was completed [6], at least 30 human genomes have been completed via NGS, and more will be available with the ongoing 1000 Genome Project [12]. Similarly, the large genomes of many pathogens, including DNA viruses, bacteria and parasites have been sequenced, and new pathogens identified via metagenomics [13], such as a new Bunyavirus sequenced from patients with unexplained fever, thrombocytopenia and leukopenia [14]. A key advantage provided by NGS in genomics is the capacity to detect low-frequency variants [15], which are important elements in both genetic and infectious diseases research [16] (see ‘NGS to study rapidly mutating viruses’ below). NGS analyses have also revealed complex scenarios, with somatic gene rearrangements in host tissues or cells being far more common than expected, and copy number variations accounting for more variation between individuals than the many recognized single nucleotide polymorphisms (SNPs) [12]. As a remarkable example, a comprehensive whole genome analysis revealed a catastrophic event – termed chromotripsis – occurring in at least 2–3% of all human cancers, whereby tens to hundreds of somatic genomic rearrangements occur with many genomic segments from distinct chromosomes reassembled in random order into a derivative chromosome [17] .

Analysis of NGS data generally requires a pipeline of bioinformatics tools, which serve a variety of purposes, including technical error correction, quality control of the data output, and ad hoc analyses relevant to specific NGS applications (e.g., variant detection, transcriptome analyses, and epigenetics). Generally, the pipeline for NGS analysis firstly includes quality control of the NGS reads by both manual and automated checking of quality scores of sequence reads; and other filtering, including elimination of the ends of individual reads which frequently carry systematic errors. After quality control, NGS reads can be aligned to a reference genome (when available), or aligned in a de novo approach (albeit with clear consequences in terms of increased complexity and analysis time). A collection of alignment tools are available – mostly public domain with many wiki pages and web sites constantly updated with the new and existing tools (e.g., see http://SEQanswers.com/). The alignment step is followed by specific data analysis for the application in consideration. Tools are also available to handle the large NGS data files and to allow for exchange in format, which facilitates the bioinformatics analyses as well as the exchange of information (e.g., samtools http://samtools.sourceforge.net/samtools.shtml). Several recent reviews have summarized currently available tools for sequence alignment [16], SNP calling [59], and transcriptomic analyses [16].

Pushed by one of the most compelling limitations of NGS, the short read length – new advanced bioinformatics tools have been developed to reconstruct long sequence haplotypes from short NGS reads. These approaches utilize advanced statistics (Bayesian and clustering algorithms), to reassemble haplotypes and estimate their frequency of occurrence in the sequence population [55]. This analysis has particular relevance to vaccine development, as variants within the pathogen population may carry sets of mutations that allow escape from host immunity, or an increase in virulence. Recent applications of NGS also allow pooling of different genomes, and analysis of genetic variants from hundreds of individuals in a single run, thereby removing a key limitation of traditional molecular genotyping with laborious and costly assays [9].

The availability of whole human genomes is also fuelling research into the diversity of haplotypes within apparently homogeneous ethnic populations, to clarify potentially the effect of variations within one gene and their interplay with other genes. [12]. Future vaccines are likely to incorporate a level of individualization based on genetic variability, such as in human histocompatibility locus antigens (HLAs), which regulate host cellular immunity by restricting antigen presentation to T cells [18]. For instance, NGS using primers tagged with an individual barcode of a few nucleotides has been used to genotype hundreds of individuals at several loci in parallel [19]. Other applications of NGS include the study of genetic variations in humans that may explain differential immune responses to the same pathogen or candidate vaccine (e.g., via functional polymorphisms in host response genes) 12, 16.

NGS and transcriptomics

As a result of its versatility and efficiency, RNA-Seq (Box 1, Table II) is rapidly becoming the gold standard technology to gather comprehensive transcriptional level information [7]. RNA-Seq has been shown to detect 25% more transcripts than microarrays [20], and has been utilized in both experimental animals and human cells [7], as well as to obtain the transcriptome of large-genome pathogens. For example, four previously unrecognized protein-coding regions, and large RNA splicing events with 229 potential donor and 132 acceptor sites, affecting 58 protein-coding genes, were revealed in human cytomegalovirus during virion production [21].

RNA-Seq also provides the capacity to study whether between-subject differences in immune responses to a pathogen or a candidate vaccine, are the result of alterations in the expression of coding genes, or whether ‘unseen’ portions of the genome are regulating the response. Recently, NGS has been used to characterize temporal changes in gene expression, at both host and pathogen level, during an infection (Table 1 ). Other NGS studies have focused on noncoding RNAs. For example, microRNAs (miRNAs) constitute a large family of small noncoding RNAs that post-transcriptionally regulate mRNAs, and thereby influence gene expression programs and hence fundamental cellular processes, with growing evidence of relevance to human disease [22]. To date, over 5000 miRNAs have been identified, including approximately 800 human miRNAs. During some viral infections, such as with herpesviruses and adenoviruses, the virus and host cell mutually cross-regulate gene expression via miRNAs [23] (see Table 1 for further examples).

Table 1

Current applications of NGS to the study of rapidly mutating viruses

Area of research	Pathogena	Refs
Detection of low frequency variants	HCVHIVSARSInfluenzaNorovirusRhinovirus	16, 4769, 70[71]49, 59[46][72]
Drug resistance	InfluenzaHBVHIVHCV	[49][73][70][74]
Host–pathogen interactions	GeneralHIVHCV	[41]39, 62, 65, 69[40]
Mechanisms of viral evolution within-host	HCVHIVRhinovirusInfluenzaHBVPoxvirus	16, 47[75][72][49][7][76]
Molecular epidemiology of pathogens	InfluenzaHIV	16, 77[16]
Detection of contaminants for vaccine safety	Poliovirus	52, 53, 54
Detection of adaptive host responses	HIV	[11]
Detection of escape variants	HIVHCV	10, 68, 69, 70, 7816, 47
Haplotype reconstruction	HIVHCVNorovirus	[79][47][46]
Detection of new strains/pathogens/genotypes[metagenomics]	Influenza virusArenavirusNorovirus/influenza	[14][80][76]

HBV, hepatitis B virus; SARS, severe acute respiratory syndrome.

NGS and epigenetic modifications

Epigenetics is the study of heritable chemical changes that occur on DNA and histone molecules, notably DNA methylation and histone deacetylation. These changes affect gene expression or cellular phenotypes, and are known to be relevant in many human diseases [24]. Epigenetic mechanisms target both host and viral genomes, and hence may have important effects on responses to viral vaccines. High-throughput technologies now allow genome-scale mapping of these modifications by adapting DNA sequencing to detect methylation sites. The first whole genome methylome of human cells was published in 2009, revealing 4–6% of the cytosine sites to be methylated [25]. A major current challenge is delineation of the scope and significance of epigenetic modifications, particularly in infectious diseases and cancer.

Chromatin immunoprecipitation (ChIP) technologies were first developed to discover DNA–protein binding sites [26]. When combined with NGS, ChIP-Seq has been applied to study how transcription factors and other chromatin-associated proteins, such as polymerases, interact with DNA to regulate gene expression. These combined technologies have already revealed a substantial component of the distribution of transcription factors involved in development of B and T cell responses [27]. In the foreseeable future, it is likely that these technologies will contribute to achieving a comprehensive understanding of the DNA-binding profiles and epigenetic modification patterns associated with immune responses to both pathogens and vaccines.

Vaccinomics, reverse vaccinology and NGS

In less than a century, the vaccines developed using Pasteur's original rules of ‘isolate, inactivate and inject the microorganism’ led to the elimination of some of the most devastating infectious diseases globally. The majority of existing vaccines were developed either using such conventional means, that is, by attenuation of the pathogen by serial passage in vitro to generate live-attenuated strains that retain immunogenicity but are no longer pathogenic; or by identification of protective antigens for use in nonliving, subunit vaccines 28, 29. For the latter, arduous and costly biochemical methods have traditionally been used to purify antigens from organisms grown in culture, resulting in small numbers of proteins ultimately being tested for immunogenicity, with limited account for existing evidence of naturally occurring protection against these antigens [30]. Since the end of the 20th century, new technologies have been proposed to address vaccines against other pathogens for which previous methods have failed. A remarkable discovery was the whole-genome sequence of H. influenzae [31], which then allowed moving beyond Pasteur's rules to investigation of pathogen genomes to inform vaccine design.

Reverse vaccinology is a relatively recent research discipline in which pathogen sequences have been utilized to predict antigenic proteins exposed on the surface of the pathogen, which can then be tested experimentally. Genome-wide sequencing has been utilized to detect potential antigenic sites in Group B meningococci – responsible for 50% of meningococcal sepsis and meningitis worldwide [31]. This bacterium had been refractory to vaccine development, because its capsular polysaccharide (polysialic acid) is nonimmunogenic (because it is expressed by several human tissues and hence is effectively a self-antigen to which the human immune system is tolerant). With the reverse vaccinology approach, 600 putative antigens were discovered, of which 29 were shown to induce antibodies that kill the bacterium in vitro via complement-mediated mechanisms. Today, five of these antigens have been inserted into a prototype vaccine (plus outer membrane vesicle component), which is completing phase III clinical trials [32]. Following this seminal application, many other pathogens are currently being targeted with the reverse vaccinology approach where previous technology have failed, such as Group A Streptococcus, Staphylococcus aureus, Streptococcus pneumoniae, and Chlamydia pneumoniae [33].

Vaccine development approaches are now taking full advantage of the explosion of high-throughput techniques [34] by utilizing genomics, transcriptomics, and proteomics (collectively termed ‘omics’) 6, 8, 27, as well as computational and statistical analysis of high throughput data (Figure 1; Box 1, Table I). NGS is thus rapidly becoming a core technology in what is now called vaccinomics 18, 30, which describes a systems biology approach to vaccine research. The ultimate goal of this approach is to delineate the cellular and molecular pathways by which pathogens induce protective immune responses, and to recapitulate, and potentially enhance, those responses via vaccination utilizing genetic signatures that predict both immunogenicity and safety, and hence efficacy (Box 1, Table II).

Bacterial pathogens and NGS

The application of NGS to study complex pathogens, such as bacteria, has significantly contributed to unraveling important elements at both the genome and transcriptome level, and has also shed light on how these pathogens evolve in response to clinical interventions 35, 36. For example, NGS has been applied to reveal the mutation and recombination events that permitted adaptation of 240 multidrug-resistant strains of S. pneumoniae [35]. More than 700 recombination events and a total of 57 736 SNPs were identified, allowing a phylogenetic reconstruction of the origin and distribution of these strains worldwide. Similarly, application of RNA-Seq to dissect the transcriptome from processed RNA of Helicobacter pylori revealed a complex scenario with the discovery of hundreds of transcriptional start sites (as opposed to the 55 previously known) [37]. This analysis also revealed more than 60 previously unknown small noncoding RNAs, that are probably involved in regulation of RNA expression and bacterial growth.

NGS to study rapidly mutating viruses

Variations in the interactions between pathogen and host determine outcomes ranging from asymptomatic infection to severe, life-threatening illness, and from efficient clearance to established chronic infection. Understanding these interactions is a key underpinning of vaccine research. This is exemplified by the application of NGS to rapidly mutating RNA viruses reviewed below.

As a result of the lack of proofreading capacity in the error-prone replicase, and to recombination events, RNA viruses mutate frequently within the host during a single infection, between hosts in a single outbreak, and across populations over time. Error rates for RNA viruses have been estimated at 10–3 to 10–5 misincorporations per nucleotide copied – almost a millionfold higher than error rates during replication of human cellular DNA. This evolutionary capacity severely limits strategies for the design of vaccines to protect populations from the large spectra of variants; well exemplified by the largely unsuccessful vaccine trials for HIV [38]. Nevertheless, application of NGS to the study of RNA viruses offers unique insights into the rapid adaptation dynamics within a single infected host, and hence the opportunity to investigate on a short time scale the role of innate and adaptive immunity during these evolutionary dynamics (Table 1).

Host–pathogen interactions revealed via NGS

NGS offers the opportunity for detailed examination of transcriptomic modifications in virus-infected host cells. For instance, a comparison of viral and host transcriptomes in HIV-infected and uninfected T cells in vitro revealed that 2.3% of the transcripts were differentially expressed, and at the peak of the infection, one in 143 transcripts was of viral origin [39]. In a study of HCV infection, RNA-Seq together with established methods (gene arrays and proteomics) provided a comprehensive description of the metabolic effects of HCV infection on target cells in vitro [40] (see also Table 1).

Another exciting application of NGS to study host–pathogen interactions is the combination of RNA hybridization and ChIP-Seq, which was recently used to study the transcriptional network of dendritic cells after stimulation with an array of pathogens [41]. This work revealed the regulatory functions of 125 transcription factors, chromatin modifiers, and RNA-binding proteins, which enabled construction of a network model consisting of 24 ‘core-regulators’ and 76 ‘fine-tuners’ that describe how pathogen-sensing pathways achieve specificity.

Understanding host immune responses

Although the complexity of adaptive immune responses is crucial to protection against pathogens, it represents a key challenge in vaccine development. NGS has recently been used to study T cell receptor (TCR) diversity, and the role of rearrangements in the VDJ (variable–diversity–joining) segments of the TCR gene in shaping the repertoire of antigen-specific T cells 10, 42, 43, 44. These analyses revealed two unexpected results: first, the vast diversity observed was even higher than previously predicted; and second, there was a substantive occurrence of identical TCR sequences between unrelated individuals. For instance, there were 10 000 complementarity-determining region (CDR)-3 sequences that were shared in naïve T cells of two non-HLA-matched individuals [44], which was unexpected given the extremely high combinatorial rearrangement potential of the VDJ region. These analyses are likely to be salient to vaccine development.

NGS has also been applied to understand better the diversity within the B cell repertoire. Notably, 14 new allelic variants in human immunoglobulin heavy chain variable region genes (IGHV) were recently identified from analyses of 108 210 human IGHV chain rearrangements from 12 individuals [45]. In HIV, studies using NGS and structural biology methods have defined in detail how the antibody, VRC01, neutralizes approximately 90% of HIV-1 strains [11]. This study also showed a vast diversity in neutralizing antibodies (NAbs) directed against autologous HIV envelope sequences across many donors, including vaccine recipients and infected individuals.

NGS to detect viral variants

A key advantage of NGS in the study of RNA viral infections is the capacity to measure the frequency of occurrence of each viral variant within a complex population. NGS has been used to detect variants at frequencies as low as 0.1% 46, 47 (Table 1). This sensitivity is crucial in vaccine research, because it allows detection of the rare resistant, or immune escape variants, which occur during natural infections (Table 1, Box 2).

In a recent investigation of within-host evolution of HCV in early (pre-seroconversion) acute infections, two potential Achilles’ heels were identified for the virus: (i) despite recognition that hundreds/thousands of viral variants are present in the inoculum from a transmission event involving shared injecting drug use apparatus, only 1–3 ‘founders’ generally established infection in the recipient; and (ii) despite a rapid and marked increase in viral diversity (more than 100 variants), which subsequently developed during early infection, a second prominent decrease in diversity was then observed within ∼100 days, associated with adaptive immune responses targeting the virus (Figure 2 ) [43]. These sequential genetic bottleneck events indicate that a potential vaccine strategy is to target the founder viruses. For HIV, it has been shown that founder viruses do have ‘phenotypic signatures’ that may be relevant for vaccine strategies, including preferential chemokine receptor, CCR5, usage and efficient replication in CD4+ T cells [48].

Figure 2

Next generation sequencing (NGS) analysis of a hepatitis C virus (HCV) population during a single acute infection showed that the virus evolved in a highly dynamic manner with strong evidence of selection pressures, which may be targeted via a preventative vaccine. (a) Phylogenetic analysis of the within-host evolution of HCV via reconstructed haplotypes of the envelope region of the genome reconstructed from NGS reads. Sequence analyses of one subject (designated 240_Ch) who ultimately developed chronic infection, revealed that the viral population found in the acute phase of the infection (aquamarine and blue, see Time legend) became markedly reduced in diversity around 100 days post-infection, before a new viral population emerged from variants that survived the genetic bottleneck event (reduction in genetic diversity) replacing the single founder virus and its progeny. Colors are also used to portray the sampling time point (see legend). This new genetically distinct viral population (gray and red in the color legend) dominated the chronic phase of infection. The size of the circles represents the prevalence of the individual variant within the viral population. (b) Kinetics of changes in viral load, and the relative contribution of individual viral variant over time are portrayed. The y axis shows the contribution of each variant with respect to the total viral RNA level. Infection was initiated with one founder variant (designated 240AF, blue line), which was then replaced sequentially by two related variants, 240AC1 and 240AC2, respectively (red unbroken and broken lines). Below the graph is a set of amino acid sequences indicating the distinguishing residues for the different variants. These sequences also show the location of a putative cytotoxic T cell (CTL) epitope (pink shading), and of antibody epitopes (green shading), as well as a mutation associated with reduction of viral reproduction from in vitro experiments (light blue shading). Figure adapted from [47].

NGS has also been applied to study the diversity of the influenza A/H1N1/2009 epidemic. For example, an immunosuppressed patient treated with neuraminidase inhibitors carried a mixed infection with three genetically distinct variants, as well as drug-resistant mutations in at least 10% of the viral population [49].

The impact of a norovirus transmission event on viral diversity, and the contribution to diversity of intrahost evolution over the typical short time period of shedding in normal hosts (10 days) or prolonged periods (up to 288 days) has recently been investigated in immunocompromized individuals. Minor variants at frequencies as low as 0.01% were successfully transmitted, indicating that transmission is an important source of norovirus diversification at the interhost level.

NGS to detect immune escape

Pathogen genome sequencing via NGS has largely removed a previous gap in vaccine development, and it is now possible to sequence the full genome of an RNA virus within days. The availability of these genomes combined with bioinformatic predictions of epitopes (see [16] for review), now allows efficient screening of pathogen genomes before experimental confirmation of immunogenicity 33, 34.

A common application of NGS is the quantification of escape variants and their frequency of occurrence at unprecedented depth and accuracy (Table 1). For instance, over 50 variant forms of each epitope in the HIV genome targeted by CD8 T cell responses during early immune escape were identified, in comparison to only 2–7 variants detected in the same samples via conventional sequencing [48]. In the context of a live-attenuated simian immunodeficiency virus (SIVmac239Δnef) vaccine administered to macaques, NGS was used to study the kinetics of occurrence of escape variants, in parallel with the evolution of the TCR β-chain repertoire specific for the wild-type epitope (Mamu-A*01 restricted Tat28–35SL8) [10]. In this interesting application of NGS to host–pathogen evolution, escape variants occurred at frequencies as low as 1% in the first 2 weeks post-vaccination, and these variants decayed rapidly in frequency over the first 8 weeks post-vaccination. Despite a diversification of the available T cell repertoire over time, the T cell response remained relatively focused on the wild type Tat28–35SL8 epitope.

Vaccine design

NGS offers the potential to improve current reverse vaccinology strategies, such as the polyvalent ‘mosaic’ HIV vaccine development. Here, in silico algorithms have been used to select viral proteins that best encompass naturally occurring HIV-1 strains [50]. These vaccines generate broad cross-reactive responses against common epitopes (see [33] for a review on reverse vaccinology strategies) and therefore offer potential for broad global application.

NGS also provides an efficient tool for surveillance of the ongoing evolution of important pathogens. This is exemplified by influenza infection, for which new vaccines have to be designed annually to account for continuing viral diversity (drift), as well as screening for major changes (shift) in incident strains potentially associated with pandemics. New influenza vaccine design therefore necessitates prompt decisions, and rapid implementation in both vaccine manufacture and field application to contain new pandemic threats [51].

Vaccine safety

NGS also has important applications in vaccine safety [52]. An NGS-based approach can be used to detect virulent mutations in vaccine batches – for instance, detection of the neurovirulent variant 472-C in the poliovirus genome from the live-attenuated polio vaccine [53]. This contrasts with standard approaches such as PCR and restriction fragment length polymorphism (PCR-RFLP) screening, which are less sensitive and limited to detection of recognized mutations. Similarly, using eight live-attenuated viral vaccines (trivalent oral poliovirus, rubella, measles, yellow fever, varicella–zoster, multivalent measles/mumps/rubella, and two rotavirus vaccines), NGS analyses have revealed minor unknown variants, as well as sequences of other viruses from the producer avian and primate tissue culture cells [54].

NGS tomorrow

In the years to come, NGS and the new third generation of single cell and single molecule sequencing [5] will become the gold standard molecular technologies in immunology, virology and vaccinology. However, there are at least three key issues for resolution in wide-scale application of NGS to vaccine research. First, the quality of NGS data must be improved by resolution of technical errors [4] (see Box 2). This may extend the current spectrum of NGS applications to the detection of more complex genetic rearrangements, such as insertion, deletion and recombination events in rapidly mutating pathogen genomes [55]. Second, in an era when vaccine research is increasingly acquiring a systems biology approach, NGS in combination with large scale and high-throughput technologies to study proteomes, such as mass spectrometry, flow cytometry (and combinations of these, such as mass cytometry that allows simultaneous measurement of >30 parameters from a single cell [56]), will provide a simultaneous, rapid, and low cost flow of integrated information on genomes, transcriptomes, and proteomes of both host and pathogen. In this scenario, computational analyses will need to be integrated into the workflow; not simply in analysis of the data, but rather as in integrated component of the study design. Finally, future developments in NGS will bring forward new challenges. For instance, single molecule third generation sequencing will probably remove the constraint of short reads, but will introduce other obstacles, such as new technical errors and challenges in experimental design.

The major targets in vaccinology are therapeutic and preventative vaccines for emerging and rapidly mutating pathogens such as HIV and HCV, as well as for complex bacterial pathogens such as Mycobacterium tuberculosis. In this context, the current focus is to understand better host–pathogen interactions. The integration of NGS with the other novel technologies described above will elucidate detailed understanding of all aspects of the virus–host interactions to guide vaccine development.