Advanced search
1 file | 114.80 MB Add to list

Revolutionizing viral and bacterial diagnostics in veterinary medicine using nanopore sequencing

Nick Vereecke (UGent)
(2023)
Author
Promoter
(UGent) , (UGent) , (UGent) and (UGent)
Organization
Abstract
Diagnostic tools in the field of medicine are required to support examination procedures as performed by practitioners. In Belgium, serological tests still represent the highest number (66%) of all tests performed in pig and cattle industry, followed by bacterial cultures (7%) and molecular assays (3%). The most important reason for this trend is the costs associated with most diagnostic tests as they do not deliver sufficient information within a single test. Even though multiplex quantitative PCRs have been designed to support detection of respiratory disease complexes in cattle and pigs or virotyping of enterotoxigenic Escherichia coli (ETEC), its cost is considered high (approximately 125 EUR) in relation to the delivered information. Over the last decades, various sequencing technologies have been released, allowing a significant reduction in sequencing cost. The availability of real-time single molecule (SMRT) nanopore sequencing (Oxford Nanopore Technologies) initiated a new era in the field of diagnostics (reviewed in Chapter 1). The use of this technology in a metagenomic workflow, allows the co-identification of viral and bacterial pathogens without prior pathogen selection. The added value of this was shown for known respiratory disease complexes in swine and cattle (PRDC and BRDC, respectively), highlighting the co-circulation of primary, secondary, and lesser-known pathogens (e.g., porcine hemagglutinating encephalomyelitis virus (PHEV) and Mycoplasmopsis arginini in PRDC and BRDC, respectively) (Chapter 3). Furthermore, it allowed (semi-)quantitative evaluation of all microbes (both viral and bacterial) to assess their relevance within the disease complex along with sequence information for viruses. The latter is an important asset as it allows viral (sub)typing which is often required to distinguish vaccines from wild type infections (e.g., porcine reproductive and respiratory syndrome virus (PRRSV)) and the subtyping of swine influenza A virus (swIAV). Obtaining these viral genome sequences from a single test also showed to be of great value in surveillance, epidemiology, and tailoring of vaccination as exemplified for porcine parvovirus type 1 (PPV1). Also, in the context of outbreaks it quickly provides sequencing information (e.g., equine herpes virus 1 (EHV-1)). While (m)any host species and sample types (including fecal, serum, respiratory, tissue, and urine) can be subjected to metagenomics, some samples require an additional targeted enrichment to deliver complete viral genomes as shown for swIAV from “dirty” oral fluids (Chapter 4). Costs associated with current tests lie within the range of multiplex PCRs, allowing its use by progressive veterinarians in livestock, companion animals, and exotics, along with various applications for academic institutions and pharmaceutical companies to date. Though, feedback from the field highlighted the need of bacterial virotyping and antimicrobial susceptibility testing (AST) to allow its extended application in routine veterinary diagnostics. This represents a tougher task since existing ad-random metagenomic workflows are not able to deliver sufficient depth on bacterial genomic information yet. Here, a spike-in approach to enrich for specific targets, while maintaining metagenomic sensitivity, is thought to guarantee its wider implementation in routine diagnostic laboratories. Therefore, fundamental knowledge on genetic virulence- and antimicrobial resistance-associated markers should be obtained. To do so, accurate and complete bacterial genomes should be available. Current databases (e.g., NCBI) do not provide enough genomic and phenotypic metadata on veterinary bacteria, thus extensive efforts were made to provide new information for relevant bacterial species. The use of long-read only data for genome assemblies was examined and performed excellent in delivering accurate and complete genomes for non-fastidiously growing bacteria such as E. coli and Actinobacillus species (Chapter 5). This is majorly due to significant improvements in raw read accuracies via new pore chemistries and base calling models over the last years. For fastidiously growing bacterial species, such as Mycoplasmopsis bovis and Brachyspira hyodysenteriae, the generation of a taxon-specific base calling model was shown of great importance to deliver Illumina-level accuracy genomes (>99.999% accuracy) (Chapter 6). Applying genome-wide association studies (GWAS) on these bacterial datasets allowed the identification of various genetic markers (genes, point mutations, and plasmids) with high diagnostic predictive power for virotyping and genetic AST (Chapter 5 and 6). Even though not all resistance phenotypes could be fully explained by these markers, epigenetics (e.g., methylation) along with transcriptomic profiling could still be addressed in the future. For the former, available raw nanopore data could be re-analyzed when more widely applicable methylation software and models are available. This would allow to provide an even broader view on the complex landscape of AMR but will also represent a (very) big challenge to get implemented in a metagenomic context. Prior to the use of these genetic markers in a “virulence and genetic AST enriched” metagenomic workflow, it is recommended to still perform extensive validation via targeted mutagenesis to deliver proof-of-concept for their actual link with virulence and AMR phenotypes (as discussed in Chapter 7). In conclusion, existing metagenomic diagnostic workflows can be enriched to detect genes associated with virulence and AST in a relatively short time frame. First fundaments for its broader application and integration in routine veterinary diagnostic laboratories can be delivered when genetic markers as described in this work will be implemented. A similar approach can be exploited to augment existing metagenomic tools for a wider variety of relevant viruses and bacteria depending on the needs within the field. In the end, this will deliver accurate all-in-one diagnostic tests which will speed-up identification and AST for relevant bacteria in acute infections and support antimicrobial drug tailorship.
Diagnostische testen zijn in de diergeneeskunde noodzakelijk om dierenartsen bij het examineren te ondersteunen. In België worden serologische testen nog steeds het vaakst (66%) toegepast in de varkens- en runderindustrie, gevolgd door bacteriële culturen (7%) en moleculaire testen (3%). De belangrijkste reden voor deze trends is de kosten die gepaard gaan met laatstgenoemde testen omdat deze vaak onvoldoende informatie leveren in één test. Ook al werden multiplex en kwantitatieve PCR’s al ontworpen om de detectie van respiratoire ziektecomplexen in runderen en varkens of pathotypering van enterotoxigene Escherichia coli (ETEC) te ondersteunen, de geassocieerde kost is nog steeds te hoog (ongeveer 125 euro per staal). De laatste jaren werden verschillende technologieën voor sequentieanalyse geïntroduceerd, wat toeliet om de kost voor het sequeneren significant te verlagen. Het beschikbaar hebben van de realtime single molecule (SMRT) nanopore technologie (Oxford Nanopore Technologies) opende een nieuw tijdperk in de diagnostiek (gereviewd in Hoofdstuk 1). Het gebruik van deze technologie in een metagenomische workflow laat toe om zowel virale als bacteriële ziekteverwekkers in een enkele test te detecteren zonder vooraf pathogenen te selecteren. De toegevoegde waarde van deze toepassing werd aangetoond voor gekende respiratoire ziektecomplexen bij varken en rund. Dit toonde aan dat er cocirculatie van primaire, secundaire, en minder gekende pathogenen (zoals porcien hemagglutinating encephalomyelitis virus (PHEV) en Mycoplasmopsis arginini respectievelijk bij varken en rund (Hoofdstuk 3)). Daarnaast laat de technologie ook toe om alle microben (zowel viraal als bacterieel) op een (semi-)kwantitatieve manier te evalueren om zo hun relevantie binnen het ziektecomplex te bepalen. Dit samen met sequentieanalyses voor virale genomen. Dat laatste is een belangrijk voordeel omdat dit het (sub)typeren van virussen toelaat, wat vaak noodzakelijk is om een onderscheid te kunnen maken tussen vaccinstammen en wild type infecties (vb. porcien reproductief en respiratoir syndroom virus (PRRSV)) en het subtyperen van influenza A-virus bij varkens (swIAV)). De beschikbaarheid van deze genoomsequenties uit één enkele test laat toe om een bijdrage te leveren bij surveillance, epidemiologie, en het eventueel op punt stellen van vaccinaties zoals voorgesteld voor het porcien parvovirus type 1 (PPV1). Ook in de context van een virale uitbraak kan het gebruikt worden om snel sequentie-informatie te bekomen (vb. equien herpes virus 1 (EHV1)). Zo goed als alle dieren en staaltypes (inclusief fecaal, serum, respiratoir, weefsel en urine) kunnen gebruikt worden voor het sequeneren in een metagenomische context. Sommige staaltypes hebben echter nood aan een extra doelgerichte verrijking om complete virale genomen te bekomen. Dit werd aangetoond voor swIAV uit “vuile” orale vloeistoffen (Hoofdstuk 4). Kosten die geassocieerd worden met huidige testen liggen in de orde van multiplex PCR’s die toelaten gebruikt te worden door progressieve dierenartsen van vee, gezelschapsdieren en exotische dieren, samen met toepassingen voor academische instituten en farmaceutische bedrijven. Feedback uit het veld suggereert dat er nood is aan verdere typering van bacteriën (vb. virotypering) en het bepalen van antimicrobiële gevoeligheid om de huidige test breder bij routinediagnostiek te kunnen toepassen. Dit houdt een grotere opgave in omdat bestaande ad random metagenomische workflows nog niet in staat zijn om voldoende diepgang te leveren voor bacteriële genomische informatie. Hierbij zou een doelgerichte spike-in benadering toelaten om een verrijking uit te voeren voor specifieke genen, met het behoud van metagenomische gevoeligheid. Dit zou een uitgebreidere toepassing bieden voor routinediagnostiek in diagnostische laboratoria. Om dit toe te laten is fundamentele kennis noodzakelijk rond genen die geassocieerd worden met virulentie en antibioticaresistentie, waarvoor accurate en complete bacteriële genomen voorhanden moeten zijn. Huidige databases, zoals NCBI, hebben vaak onvoldoende genomische en fenotypische metadata voor veterinaire bacteriën beschikbaar. Dus werd moeite gedaan om deze informatie te voorzien voor relevante bacteriële species. Het gebruik van lange reads voor genoomassemblage werd uitgebreid getest en kon voor “standaard” bacteriën zoals E. coli en Actinobacillus species excellente genomen met voldoende accuraatheid en compleetheid afleveren (Hoofdstuk 5). Dit is vooral mogelijk door de significante verbetering in accuraatheid van ruwe reads via nieuwe technologie-updates waaronder nanopore chemie en base calling algortimes over de laatste jaren heen. Voor “niet-standaard” bacteriële species, zoals Mycoplasmopsis bovis en Brachyspira hyodysenteriae, was een taxon-specifiek base calling model noodzakelijk om gouden standaardniveau genomen (Illumina; >99.999% accuraatheid) af te leveren (Hoofdstuk 6). Het toepassen van genoom-wijde associatiestudies (GWAS) op deze bacteriële datasets liet toe om verscheidene genetische merkers (genen, puntmutaties en plasmides) te identificeren met hoge diagnostische voorspellende kracht voor zowel virotypering en bepaling van antibioticagevoeligheid (Hoofdstuk 5 en 6). Ook al konden niet alle resistentie fenotypes compleet voorspeld worden, in de toekomst kunnen deze via epigenetica (vb. methylatie) of het profileren van het bacterieel transcriptoom verder onderzocht worden. Voor die eerste kan de beschikbare ruwe nanopore data opnieuw geanalyseerd worden als er meer wijder toepasbare methylatie software en modellen voorhanden zijn. Dit zou toelaten om het complexe landschap van antibioticaresistentie breder te bestuderen. Het zal een (zeer) grote uitdaging zijn om deze data in een metagenomische context en workflow te implementeren. Vooraleer deze genetische merkers in een “virulentie en genetische gevoeligheidsbepaling aangereikte” metagenomische workflow kunnen gebruikt worden, is het aangeraden om nog uitgebreide validaties uit te voeren via doelgerichte mutagenese om zo proof-of-concept te bieden voor de bijdrage van deze merkers tot het virulentie en/of resistentie fenotype (zoals besproken in Hoofdstuk 7). Conclusie: in een relatief korte tijd kunnen metagenomische diagnostische workflows aangereikt worden met genetische markers, die geassocieerd worden met virulentie en antibioticaresistentie. Deze eerste fundamenten kunnen worden getest door de genetische informatie – zoals beschreven in dit werk – te implementeren om zo een breder toepassingsveld en integratie in veterinaire diagnostische laboratoria te bekomen. Een vergelijkbare benadering kan gebruikt worden om bestaande metagenomische workflows op punt te stellen voor een wijdere variëteit aan relevante virussen en bacteriën in het veld. Finaal zal dit ertoe leiden dat bij acute infecties accurate all-in-one diagnostische tests identificatie, typering en genetische antibioticaresistentiebepaling voor relevante bacteriën kunnen versnellen om zo het gebruik van antibiotica beter te sturen.
Keywords
infectious diseases, nanopore sequencing, viruses, bacteria, diagnostics

Downloads

  • PhD Manuscript NickVereecke withCover.pdf
    • full text (Published version)
    • |
    • open access
    • |
    • PDF
    • |
    • 114.80 MB

Citation

Please use this url to cite or link to this publication:

MLA
Vereecke, Nick. Revolutionizing Viral and Bacterial Diagnostics in Veterinary Medicine Using Nanopore Sequencing. Ghent University. Faculty of Veterinary Medicine, 2023.
APA
Vereecke, N. (2023). Revolutionizing viral and bacterial diagnostics in veterinary medicine using nanopore sequencing. Ghent University. Faculty of Veterinary Medicine, Ghent, Belgium.
Chicago author-date
Vereecke, Nick. 2023. “Revolutionizing Viral and Bacterial Diagnostics in Veterinary Medicine Using Nanopore Sequencing.” Ghent, Belgium: Ghent University. Faculty of Veterinary Medicine.
Chicago author-date (all authors)
Vereecke, Nick. 2023. “Revolutionizing Viral and Bacterial Diagnostics in Veterinary Medicine Using Nanopore Sequencing.” Ghent, Belgium: Ghent University. Faculty of Veterinary Medicine.
Vancouver
1.
Vereecke N. Revolutionizing viral and bacterial diagnostics in veterinary medicine using nanopore sequencing. [Ghent, Belgium]: Ghent University. Faculty of Veterinary Medicine; 2023.
IEEE
[1]
N. Vereecke, “Revolutionizing viral and bacterial diagnostics in veterinary medicine using nanopore sequencing,” Ghent University. Faculty of Veterinary Medicine, Ghent, Belgium, 2023.
@phdthesis{01GVJ5E65QE5AAQ6DB6G8DX3CD,
  abstract     = {{Diagnostic tools in the field of medicine are required to support examination procedures as performed by practitioners. In Belgium, serological tests still represent the highest number (66%) of all tests performed in pig and cattle industry, followed by bacterial cultures (7%) and molecular assays (3%). The most important reason for this trend is the costs associated with most diagnostic tests as they do not deliver sufficient information within a single test. Even though multiplex quantitative PCRs have been designed to support detection of respiratory disease complexes in cattle and pigs or virotyping of enterotoxigenic Escherichia coli (ETEC), its cost is considered high (approximately 125 EUR) in relation to the delivered information. Over the last decades, various sequencing technologies have been released, allowing a significant reduction in sequencing cost. The availability of real-time single molecule (SMRT) nanopore sequencing (Oxford Nanopore Technologies) initiated a new era in the field of diagnostics (reviewed in Chapter 1). 
The use of this technology in a metagenomic workflow, allows the co-identification of viral and bacterial pathogens without prior pathogen selection. The added value of this was shown for known respiratory disease complexes in swine and cattle (PRDC and BRDC, respectively), highlighting the co-circulation of primary, secondary, and lesser-known pathogens (e.g., porcine hemagglutinating encephalomyelitis virus (PHEV) and Mycoplasmopsis arginini in PRDC and BRDC, respectively) (Chapter 3). Furthermore, it allowed (semi-)quantitative evaluation of all microbes (both viral and bacterial) to assess their relevance within the disease complex along with sequence information for viruses. The latter is an important asset as it allows viral (sub)typing which is often required to distinguish vaccines from wild type infections (e.g., porcine reproductive and respiratory syndrome virus (PRRSV)) and the subtyping of swine influenza A virus (swIAV). Obtaining these viral genome sequences from a single test also showed to be of great value in surveillance, epidemiology, and tailoring of vaccination as exemplified for porcine parvovirus type 1 (PPV1). Also, in the context of outbreaks it quickly provides sequencing information (e.g., equine herpes virus 1 (EHV-1)). While (m)any host species and sample types (including fecal, serum, respiratory, tissue, and urine) can be subjected to metagenomics, some samples require an additional targeted enrichment to deliver complete viral genomes as shown for swIAV from “dirty” oral fluids (Chapter 4). 
Costs associated with current tests lie within the range of multiplex PCRs, allowing its use by progressive veterinarians in livestock, companion animals, and exotics, along with various applications for academic institutions and pharmaceutical companies to date. Though, feedback from the field highlighted the need of bacterial virotyping and antimicrobial susceptibility testing (AST) to allow its extended application in routine veterinary diagnostics. This represents a tougher task since existing ad-random metagenomic workflows are not able to deliver sufficient depth on bacterial genomic information yet. Here, a spike-in approach to enrich for specific targets, while maintaining metagenomic sensitivity, is thought to guarantee its wider implementation in routine diagnostic laboratories. Therefore, fundamental knowledge on genetic virulence- and antimicrobial resistance-associated markers should be obtained. To do so, accurate and complete bacterial genomes should be available. Current databases (e.g., NCBI) do not provide enough genomic and phenotypic metadata on veterinary bacteria, thus extensive efforts were made to provide new information for relevant bacterial species. The use of long-read only data for genome assemblies was examined and performed excellent in delivering accurate and complete genomes for non-fastidiously growing bacteria such as E. coli and Actinobacillus species (Chapter 5). This is majorly due to significant improvements in raw read accuracies via new pore chemistries and base calling models over the last years. For fastidiously growing bacterial species, such as Mycoplasmopsis bovis and Brachyspira hyodysenteriae, the generation of a taxon-specific base calling model was shown of great importance to deliver Illumina-level accuracy genomes (>99.999% accuracy) (Chapter 6). Applying genome-wide association studies (GWAS) on these bacterial datasets allowed the identification of various genetic markers (genes, point mutations, and plasmids) with high diagnostic predictive power for virotyping and genetic AST (Chapter 5 and 6). Even though not all resistance phenotypes could be fully explained by these markers, epigenetics (e.g., methylation) along with transcriptomic profiling could still be addressed in the future. For the former, available raw nanopore data could be re-analyzed when more widely applicable methylation software and models are available. This would allow to provide an even broader view on the complex landscape of AMR but will also represent a (very) big challenge to get implemented in a metagenomic context. Prior to the use of these genetic markers in a “virulence and genetic AST enriched” metagenomic workflow, it is recommended to still perform extensive validation via targeted mutagenesis to deliver proof-of-concept for their actual link with virulence and AMR phenotypes (as discussed in Chapter 7).
In conclusion, existing metagenomic diagnostic workflows can be enriched to detect genes associated with virulence and AST in a relatively short time frame. First fundaments for its broader application and integration in routine veterinary diagnostic laboratories can be delivered when genetic markers as described in this work will be implemented. A similar approach can be exploited to augment existing metagenomic tools for a wider variety of relevant viruses and bacteria depending on the needs within the field. In the end, this will deliver accurate all-in-one diagnostic tests which will speed-up identification and AST for relevant bacteria in acute infections and support antimicrobial drug tailorship.}},
  author       = {{Vereecke, Nick}},
  keywords     = {{infectious diseases,nanopore sequencing,viruses,bacteria,diagnostics}},
  language     = {{eng}},
  pages        = {{XIV, 318}},
  publisher    = {{Ghent University. Faculty of Veterinary Medicine}},
  school       = {{Ghent University}},
  title        = {{Revolutionizing viral and bacterial diagnostics in veterinary medicine using nanopore sequencing}},
  year         = {{2023}},
}