ContentslistsavailableatScienceDirect
Journal of Infection
journalhomepage:www.elsevier.com/locate/jinf
The global meningitis genome partnership
Elizabeth Rodgers
a,∗, Stephen D. Bentley
b, Ray Borrow
c, Holly B. Bratcher
d, Sylvain Brisse
e, Angela B. Brueggemann
f, Dominique A. Caugant
g, Jamie Findlow
h, LeAnne Fox
i,
Linda Glennie
a, Lee H. Harrison
j, Odile B. Harrison
d, Robert S. Heyderman
k, Melissa Jansen van Rensburg
f, Keith A. Jolley
d, Brenda Kwambana-Adams
k,
Shamez Ladhani
l, Marc LaForce
m, Michael Levin
n, Jay Lucidarme
c, Neil MacAlasdair
b, Jenny Maclennan
d, Martin C.J. Maiden
d, Laura Maynard-Smith
d, Alessandro Muzzi
o, Philipp Oster
p, Charlene M.C. Rodrigues
d, Olivier Ronveaux
q, Laura Serino
o, Vinny Smith
a, Arie van der Ende
r, Julio Vázquez
s, Xin Wang
i, Saber Yezli
t, James M. Stuart
qaMeningitis Research Foundation, Newminster House, 27-29 Newminster House, Baldwin Street, Bristol BS1 1LT, UK
bWellcome Sanger Institute, Parasites and microbes, Hinxton CB10 1SA, UK
cPublic Health England, Meningococcal Reference Unit, Manchester Royal Infirmary, Manchester M13 9WZ, UK
dDepartment of Zoology, University of Oxford, Oxford OX1 3SY, UK
eInstitut Pasteur, Biodiversity and Epidemiology of Bacterial Pathogens, Paris, France
fNuffield Department of Population Health, University of Oxford, Oxford OX3 7LF, UK
gDivision for Infection Control and Environmental Health, Norwegian Institute of Public Health, Oslo, Norway
hPfizer Limited, Walton Oaks, Dorking Road, Tadworth, Surrey KT20 7NS, UK
iMeningitis and Vaccine Preventable Disease Branch, Division of Bacterial Diseases, Centers for Disease Control and Prevention, United States
jInfectious Diseases Epidemiology Research Unit, University of Pittsburgh, Pittsburgh, PA, United States
kNIHR Global Health Mucosal Pathogens Research Unit, Division of Infection & Immunity, University College London, London, UK
lPublic Health England, Immunisation and Countermeasures Division, 61 Colindale Avenue, London NW9 5EQ, UK; Paediatric Infectious Diseases Research Group (PIDRG), St. George’s University of London, Cranmer Terrace, London SW17 0RE, UK
mSerum Institute of India, Ltd, Pune, India
nImperial College London, Paediatrics, London, UK
oGSK, Siena, Italy
pSanofi Pasteur, Lyon, France
qWHO, Geneva, Switzerland
rDepartment of Medical Microbiology and Infection Prevention, University of Amsterdam, Amsterdam UMC and, the Netherlands Reference Laboratory for Bacterial Meningitis, Amsterdam, the Netherlands
sInstitute of Health Carlos III, Madrid, Spain
tMinistry of Health, The Global Centre for Mass Gatherings Medicine, Riyadh, Saudi Arabia
a rt i c l e i nf o
Article history:
Accepted 26 June 2020 Available online 29 June 2020 Keywords:
Whole genome sequencing Genome partnership Neisseria meningitidis Streptococcus pneumoniae Streptococcus agalactiae Haemophilus influenzae Bacterial meningitis Epidemiology
s u m m a ry
Genomicsurveillanceofbacterialmeningitispathogensisessentialforeffectivediseasecontrolglobally, enablingidentification ofemergingand expandingstrains andconsequent publichealthinterventions.
Whiletherehasbeenariseintheuseofwholegenomesequencing,thishasbeendrivenpredominately byasubsetofcountrieswithadequatecapacityandresources.Globalcapacitytoparticipateinsurveil- lanceneedstobeexpanded,particularlyinlowandmiddle-incomecountrieswithhighdiseaseburdens.
Inlight ofthis,the WHO-ledcollaboration,DefeatingMeningitisby 2030GlobalRoadmap, hascalled fortheestablishmentofaGlobalMeningitisGenomePartnershipthatlinksresourcesfor:N.meningitidis (Nm),S.pneumoniae(Sp),H.influenzae(Hi)andS.agalactiae(Sa)toimproveworldwideco-ordinationof strainidentificationandtracking.Existingplatformscontainingrelevantgenomesinclude:PubMLST:Nm (31,622),Sp(15,132),Hi(1935), Sa(9026);TheWellcomeSangerInstitute:Nm(13,711),Sp(> 24,000), Sa(6200),Hi(1738);andBMGAP:Nm(8785),Hi(2030).Asteeringgroupisbeingestablishedtocoor- dinatetheinitiativeandencouragehigh-qualitydatacuration.Nextstepsinclude:developingguidelines onopen-accesssharingofgenomicdata;definingacoresetofmetadata;andfacilitatingdevelopmentof user-friendlyinterfacesthatrepresentpubliclyavailabledata.
© 2020TheAuthors.PublishedbyElsevierLtdonbehalfofTheBritishInfectionAssociation.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense.
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
https://doi.org/10.1016/j.jinf.2020.06.064
0163-4453/© 2020 The Authors. Published by Elsevier Ltd on behalf of The British Infection Association. This is an open access article under the CC BY-NC-ND license.
( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Introduction
Meningitis,andrelatedinvasivebacterialinfectionssuchasbac- teraemicpneumoniaandsepsis,aredevastatingdiseasesthatrep- resenta majorglobalhealthchallenge.1,2 Whilesuccessfulimmu- nisation programmes against three of the main bacterial causes, Neisseriameningitidis(Nm,themeningococcus),Streptococcuspneu- moniae (Sp, the pneumococcus), and Haemophilus influenzae type b (Hib), have enabled great progress to be made, the bur- denremainshighwithprogresssubstantially behindthatofother infectious diseases.According toone estimate,between1990 and 2017 child (under five) meningitis deaths fell by just 53%, com- paredto70%fordiarrhoea,87%formeasles,and93%fortetanus.3 Inthe Africanmeningitisbelt, largeepidemicsofmeningococ- cal diseaseoccur periodically.Although therehavebeendramatic declines in Nm serogroup A (NmA) disease following successful MenAfriVac® mass vaccinationcampaigns,the threat ofboth en- demicandepidemicdiseasecausedbynon-ANmandSppersists4, withAfricaasawholestillexperiencingthegreatestglobalburden ofbacterialmeningitis.Epidemiologicaldataunderlyingglobalbur- den estimates,particularly inAsia, remain incomplete,however.5 Nucleotidesequence-basedtyping,inparticularwholegenomese- quencing (WGS), hasimportant applications fortracking and re- spondingtoendemicandepidemicdisease.
WGStoelucidatetheoriginandspreadofnewstrainsfor publichealthmanagement
Outbreaks ofNm serogroup C(NmC) disease wererare inthe meningitis belt until 2013, when North-Western Nigeria experi- enced localisedoutbreaks causedby a novelNmC strain (the ST- 10217 clonal complex; cc10217).6 In 2015, this strain caused a severe epidemic in Niger7 and, in 2017, the largest ever known outbreak ofNmC disease in Nigeria,with over 14,000 suspected cases.8,9 The analysis and comparison of WGS data successfully elucidatedthe originandspreadofthisnewstrain that emerged after the acquisition of virulence genes by a non-encapsulated, non-virulent ancestor.10 Metagenomic approaches have since re- vealed that NmCcc10217 hasspread outsidethemeningitis belt, causinganoutbreakassociatedwithafuneralinLiberiain2017.11
GenomicapproacheswerealsoabletoresolveNmserogroupW (NmW) ST-11 clonalcomplex (cc11) isolates into multiplestrains within two divergent sublineages: (i) the ‘Hajj-strain sublineage’, (named after a constituent strain that caused a global outbreak among Hajj pilgrims in the early 2000s); and, (ii) the ‘South American-strainsublineage’thatrecentlyspreadfromSouthAmer- ica to Europeand beyond.12 Identification ofthe emergence and expansion of theSouth American-strain sublineage inthe United Kingdom (U.K.)led tointroduction ofMenACWY conjugatevacci- nationfor adolescents.A novel,seemingly morevirulent, variant, the ‘2013-strain’ hassince expanded intoEurope, Australasia and NorthAmerica13withconsequentialvaccinepolicychangesinIre- land,theNetherlands,SwitzerlandandAustralia.
WGS also has important applications in the development of vaccines to control endemic disease. Indeed the identification of therecombinantantigensusedinBexsero®, aprotein-basedvac- cinethat targetsNmserogroup B(NmB)disease,markedthefirst use ofgenomics for vaccineantigen discovery througha ‘reverse vaccinology’approach.14Priortothis,broadlyprotectiveMenBvac- cinedevelopmentwashinderedbythefactthatNmBpolysaccha- rideispoorly immunogenicandhaspossibleautoimmuneeffects, duetoitssimilaritieswithhumanglycoproteins15,whilecandidate
∗ Corresponding author.
E-mail address: [email protected] (E. Rodgers).
outer-membrane vesicle vaccines offered limited protection. Two recombinantproteinMenBvaccines,Trumenba® andBexsero® are currentlylicensed for use in multiple countries worldwide, with the latter routinely used in the U.K., Ireland, Lithuania, Andorra, SanMarinoandSouthAustralia16,whileregionsofItalyroutinely usebothTrumenba® and Bexsero®. Whilethesevaccines areex- pected to providebroad coverage of circulatingNmB strains17,18, on-going global genomic surveillance is required to provide evi- denceofvaccine impact, documentstrain coverage, monitor vac- cineescapeandtofacilitate developmentofnext generation vac- cines.19,20
WGShasan importantrole toplayinthesurveillanceofinva- sive pneumococcal disease (IPD),which includesmeningitis, bac- teraemic pneumonia, and sepsis and remains a leading global causeofmortality,especiallyamongchildren.21–23Underimmune selectivepressure24 the pneumococcus can escape pneumococcal conjugatevaccine(PCV)controlthroughserotypereplacement.The increase of non-vaccine serotypes, concomitant with a decrease in vaccine serotypes,25 highlights the importance of monitoring serotypescirculatingpre-andpost-PCV introductiontoenablein- vasive non-vaccine pneumococcal serotypes to be rapidly identi- fied.17WGShasenabledfargreaterinsightintothisprocess.
Genomicshasbeenusedtoinvestigateserotypereplacementaf- tertheintroductionofboththeoriginal(PCV7)andhighervalency (PCV10andPCV13)vaccines, andemerging non-vaccineserotypes have been found to vary extensively among countries. Using genome-wide variation, an international definition of pneumo- coccalpopulation structure ‘Global Pneumococcal Sequence Clus- ters (GPSCs)’ has been developed and enables the production of pneumococcaldatasets,independentofpreviousdefinitionsoflin- eages.26 Using this cluster definition, it wasrevealed that, post- PCV13 introduction, the top emerging cluster, GPSC3, expressed serotype8inSouthAfrica,but33FinIsraelandtheUnitedStates (U.S.).27 This demonstrateshow species-wide approachescan en- abletheevolution ofthepneumococcus tobe bettertracked,be- yondthelimitsofserotype.
Atthetime ofwriting, comprehensivegenomestudiesinvesti- gating the meningococcusand thepneumococcus were available, butsuchstudies forH. influenzaewere morelimited. It isimpor- tantto ensurethat thispathogenisadequatelyexamined, as,de- spitedramatic reductionsin Hib disease followingglobal vaccine implementation, with 192 countries having introduced a univer- salHibvaccinationprogrammeby201928,H.influenzaemeningitis causedbyother typesstilloccurs.29 Forexample,increasing rates ofinvasivediseaseduetoH.influenzaetype a(Hia)havebeende- scribedinindigenouspopulations innorthernCanada,Alaska,the southwesternU.S.andAustralia,withan additionalstudyindicat- ing that H. influenzae strainsare able to capsule-switch or mod- ulate the expression of the type b capsule.30–32 Unencapsulated non-typeableH.influenzae strains(NTHi)arealsoincreasingly as- sociated withinvasive disease, includingbacteraemic pneumonia.
Theseorganismsrepresentagrowing challengeinhealthcare set- tings,particularlyasmanyareantibioticresistant.33
Asof mid-2020 Streptococcusagalactiae(GBS),a leadingcause ofneonatalmeningitisandsepsis,wasnotyetvaccine-preventable, althoughvaccines were indevelopment.34–36 Anincreasing num- ber of WGS studies investigating GBS were becoming available, includingsome fromlow income countries, indicating a growing awarenessofthisinfectionalongsidethebenefits ofWGSdatain understandingtheepidemiologyofthispathogen.37
Consequently, continued surveillance and characterisation of bacteriaassociatedwithmeningitisremainapriority,iftheglobal burdenofmeningitisistobereduced.Theavailabilityofsequence- baseddatawillplayanessentialroleinstrengtheningsurveillance ofdiseasetypesandenablinginterventionstobeappropriatelytar- geted.
Strengtheningglobalgenomicsurveillance:aglobalmeningitis genomepartnership
The needforaGlobal MeningitisGenomeLibrary(GMGL)was identifiedinMay2017atameetingofinternationalexpertshosted by Wilton Park,38 and led by Meningitis Research Foundation (MRF)incollaborationwiththeWorldHealthOrganization(WHO).
This meeting first outlined a global vision to defeat meningitis andcalledforanactionplaninlinewiththeUnitedNationsSus- tainableDevelopmentGoals(SDGs).39,40 Subsequentconsultations ledto the establishmentofa Global Meningitis Genome Partner- ship(GMGP)tolinkresourcesforthefourleadingcausesofacute bacterial meningitis, with the aim of improving worldwide co- ordinationof strain identification andtracking, and enablingthe publichealthbenefitsofWGStobedelivered.
A Technical Task Force41 co-ordinated by WHO developed a GlobalRoadmaptoDefeatMeningitisby2030, whichwillbesub- mitted forconsideration at the World Health Assembly in 2020.
Theroadmapfocusesonthemaincausesofacutebacterialmenin- gitis that are alreadyvaccine-preventable or maybe in the fore- seeablefuture:N.meningitidis, S.pneumoniae,H. influenzae,andS.
agalactiae.Sequence-baseddataareonekeyelementtotheglobal vision to defeat meningitis, and therefore the importance ofthe establishment of a GMGP functional for the four pathogens has beenhighlightedbythe TaskForce.The terminology‘partnership’
(withinwhichlibraries, ordata repositoriesandplatforms play a crucial part), reflects the collaborative approach required forthe workthatneedstobedone.
Thisarticleprovidesanaccountofthecurrentstatusofmenin- gitis genomics globally and, in particular, focuses on: (i) review- ing existing platforms that facilitate the submission and analysis ofWGSdata;(ii)describing theeffortsofexisting globalprojects andpartnersfocusedonenhancingsurveillanceandincreasingour understandingofthesebacteria,particularlyinresource-poorset- tings; and, (iii) discussing the complexities associated with en- suringgloballyrepresentative datasetscanbe obtainedandmade publiclyavailable.
Examplesofexistingmeningitispathogengenomelibrariesand analysisplatforms
PubMLST
The open-accessPubMLST.orgwebsiteanditsunderlyingweb- based software platform, the Bacterial Isolate Genome Sequence Database (BIGSdb), include all levels of sequence data enabling ahierarchical approachto studying microbialspecies fromsingle genefragments,to functional genes groupedinto schemes,up to fullgenomecomparisons.42Theintegrationofisolatecharacterisa- tion,linkinggenotypesandgeneticvariationwithprovenanceand phenotype data, addresses a wide range of functional questions, frompopulationbiologytoantimicrobialresistance(AMR)andvac- cineformulation,allofwhichcaninformpublichealthpolicies.
At the time of writing the PubMLST databases hosted over 100 species and/or genera, including curated databases for the fourmeningitispathogens. Thisincluded> 30,000N. meningitidis genomes,> 15,000 S.pneumoniae genomes, > 9000 S.agalactiae genomesand1900H. influenzaegenomes, predominantlysubmit- tedfromhigh-incomeregionsorinternationalcentres.
PubMLSTisbasedongene-by-genepopulationconsensusanno- tation, where genes are identified and their variation catalogued systematically.43 As a result, the genetic variation of every gene inevery bacterialgenomecan,inprinciple,belinked withprove- nanceandphenotypicinformation.44Userscanperformasuiteof analyses,including:evolutionaryanalyses,bacterialspeciesclassi- fication, disease surveillance, phylogeography and outbreak anal-
yses. The provision of a common nomenclature, i.e. a universal language to describe bacterial strains and lineages,is integral to itssuccess,enablingportable useofthisinfrastructureandallow- inggloballaboratoriesandpublichealthagenciestocompareiso- latesofinterest,includingfoodbornepathogens.Asaresult,enteric pathogenoutbreakshavebeendetectedandhaltedthroughquick recognitionofthesamestrainfromdifferentsources.45–47
Inadditionto providingcuratedgenomes,visual analytictools enable users to assess WGS genetic relatedness with integrated, third-party software, such as: (i) GrapeTree48 and/or PhyloViz49; (ii)InteractiveTreeofLife(iTOL)50;and(iii)MicroReact,whichre- spectivelyenablethecreationofminimumspanningtrees,phylo- genetic trees and the visualisationof genomic epidemiology and phylogeography(Fig.1).51
The PubMLST.org Neisseria site is one of the largest single databaseswithinPubMLST(Fig.2).52Dataareactivelycuratedwith genomessubmitted byusersinternationally.Additionally,curators sometimes assemblegenomes fromshortread data fromtheEu- ropean Nucleotide Archive (ENA) and associate these with pub- lishedmetadata.Curateddataaremadefreelyavailableandacces- sibleviathewebsiteorapplicationprogramminginterfaces(APIs).
This enables third party analysis sites to interconnect with the database, facilitating open sharing of publicly available genomic data. Through APIs, comparisons can be made between private data,containingclinicallysensitiveinformation,andpubliclyavail- abledata,enablingindividualdatabasestoactsynergistically.53
The PubMLST Neisseria database containsseveral datasets, in- cluding comprehensive meningococcal genome libraries for Eng- landandWales(July2010-present),54 Scotland(2009-present)and Ireland (2010–2014); however, it is not yet globally representa- tive.AsofMarch2020,127countrieshadsubmittedmeningococ- calisolateswithsequencedatabutonly97countrieshadsubmit- tedmeningococcalgenomes andthenumberofgenomessubmit- ted(n=31,622)wasapproximatelyhalfofthe numberofisolates (n=60,776); theproactive collectionand curationof comprehen- siveandrepresentativecountry-specificgenomedatafromaround theworldcouldremedythis.
ThePubMLST.org S.pneumoniaedatabasecontainsover15,000 genomes of which over 9000 were included in the first version of the Pneumococcal Genome Library. The library aims to pro- videuserswithassembledpneumococcalgenomedataandcorre- sponding metadata, as well as an assessment ofgenome quality amongpublisheddata.Itwasdevelopedfollowinga reviewofar- ticles withatleastone pneumococcal genome; 979articles were screened on PubMed, and 173 peer-reviewed publications iden- tified, of which 97 (56%) provided retrievable pneumococcal ge- nomicdatawithoutdiscrepancies(e.g.missingorincorrectgenome sequence accession numbers).Upon firstrelease,the librarycon- taineddata from92 publications,from41 countriescollectedbe- tween1916and2018,representing>90serotypes.Effortsareun- derway toresolve discrepancies inthe remaining 76 publications andenableuploadoftheassociatedgenomes.
BIGSdbisalsousedbytheInstitutPasteur,Paristoruntheir18 databases of MLST and genome-based typing schemes. The plat- formprovidesreferencenomenclatures formicrobial isolates,pri- marilyintendedformolecularepidemiologyofpathogensofpublic healthimportance,detectionofvirulence andAMR genes,andfor populationbiologyresearch.
TheWellcomeSangerInstitute
TheWellcomeSangerInstitutehasactiveresearchforanumber of meningitis-causing pathogens.55 Datasets representbroad geo- graphicandtemporalrangesthatareofvalueforevolutionaryand epidemiologicalanalyses,butalsoincludedenselysampledcollec- tionsthatelucidatepatternsofpathogentransmissionandspread.
Fig. 1. Phylogeographic, temporal, and phylogenic visualisation tool Microreact available through the PubMLST.org website. Collection of 107 characterised meningococcal isolates chosen to be representative of disease globally in the latter half of the 20th century.
Fig. 2. PubMLST Neisseria database: isolate and genome submissions over time.
Databases ofall publicly available genomes are compiled andin- frastructureisbeingdevelopedtoenablepublicaccesstodataand analysistools.
AdatacollectionforNeisseriagenomeanalyseshasbeenestab- lishedthat gathers allpublicly available Neisseria WGSdata from theENA andallavailable associated metadata,accessedfromthe PubMLST.orgNeisseria database andthe publishedliterature. This hasgenerateda collectionofgenomes from32,538 isolates, from 77countries,collectedbetween1937and2014.Ofthese,16,164are Nmisolates,withmetadataavailablefor13,711.However,asignifi- cantproportionoftheNmisolatesarefromtheU.K.(n=4256)and otherresource-rich countries;thereforesome effortisstill neces- sarytoestablishtrulyglobalcollectionsofNeisseriaisolates.
The Sanger pneumococcal database includes the Global Pneu- mococcalSequencing (GPS)project,whichaimsto assesstherisk ofpneumococcalvaccineevasionbysamplingpopulationgenomic evolutionbefore andafter vaccine implementation,to informfu- turevaccine andsurveillance strategies.56,57 The GPS project has generatedover24,000high-qualitygenomes(3105frommeningitis cases)from58 countries,withan emphasis onlow- andmiddle- incomecountries(LMICs)wherediseaseburdenishighest.Exten- sivemetadatafor13,454ofthegenomesandinteractivevisualisa- tionsareavailableviaPneumogen.net58andPathogenwatch.59
AglobalgenomicsurveyofGBS(‘JUNO’)isnowunderway,aim- ingto provide dataandinsights to informvaccinestrategies and elucidate the routesof infection leading to infantmortality. Like theGPS project, itwill focus on trainingpartnersfromLMICs to increaseaccesstogenomedataanalysis, and26foundingpartner institutionsfrom12 countries(mostly Africancountries)have al- readyagreedtoparticipateinthisinitiative.
TheBacterialMeningitisGenomeAnalysisPlatform
Inaddition toglobalgenomecollections,platforms customised forcountryspecificneedshavebeendeveloped,includingtheU.S.
CentersforDiseaseControlandPrevention(CDC)BacterialMenin- gitisGenomeAnalysisPlatform(BMGAP).BMGAPisanautomated genomicanalysisplatformfocusedon H.influenzaeandN.menin- gitidisthatstreamlinesworkflowsandreducesanalysistime.
BMGAP relies on PubMLST nomenclature and many functions overlapwithPubMLST;however,clinical dataaccessistightly re- stricted,requiringstatepermission,andaccesstoBMGAPrequires userregistration.DuetothelowburdenofdiseaseintheUS,even anonymiseddatamaybe traceable toan individual ifprecisede- tailsoftime,place,andclinicaldataarepublishedwithgenomes;
therefore,clinicaldatasharingiscloselyregulated,oftenrequiring additional approval. However, U.S. sequencing data from BMGAP can be made publicly available in PubMLST, in compliance with CDCdata-sharingpolicies,withstrictlylimitedaccesstosensitive metadataandepidemiologicaldata,toenablethesystemstocross- talk.
Enhancingcountryrepresentationandaccessthroughglobal partnersupport
Current bacterial meningitis genome repositories are over- representativeofthe locationswhereusers alreadyhavecapacity andaccess to sample collections and understand the benefits of WGS,orwhereusersarecollaboratingwiththosewhocanobtain, cultureandsequencepatientsamples.Asaresult,sequencesfrom countriesimportant in the emergence and spreadof strains, but wholack capacity,resources, and/or funding,or who arenot au- thorisedtosharedatawithnewpartnersorcollaborators,maybe absentfromcurrentcollections.
Todevelopgloballyrepresentativegenomerepositories,aproac- tiveapproachwhichutilises partnerengagementisneeded.WHO
iscommittedtohelpingpromotethebenefitsofWGSthroughre- gional andcountry links, andthrough existing networks such as theInvasiveBacterialVaccine-PreventableDiseases(IB-VPD)global surveillancenetwork.60
TheGlobalMeningococcalInitiative(GMI)61alsohasanimpor- tantrole toplayinpromoting thebenefitsofWGSandparticipa- tionin aGMGPamongits collaborators.TheGMI isa multidisci- plinary expert group, withmembers across 37 countries. It aims topreventinvasive meningococcaldiseaseworldwidethroughed- ucation, research, andinternationalcollaboration. The role of the GMI is particularly important in regions such asEastern Europe, Asia and the Pacific where formal meningitis surveillance is less established.62 A GMI meeting in China fostereda China-U.K. col- laborationforthecharacterisationofacollectionofChineseNmW isolates63 thatwereshowntobedistinctfromtheHajj-andSouth American-strainsublineagespreviouslydescribed.64Thecollabora- tionfurtherdocumentedthedisappearanceofNmAandtheemer- gence of new NmB clones. The GMI is currently engaging with EasternEuropean countriestoinvestigatethepersistence ofNmA diseaseinKazakhstan.65
A further approach to engaging with countries not currently submitting to WGS repositories is through funding time-limited projects that establish partnerships betweencountries that have capacity and expertise in DNA extraction and sequencing, with those that do not. Data ownership and control over the level of metadata shared would reside with the corresponding country, owingto thecodes ofpracticeto beestablished regardinguseof repositories andpublication of data. Projectssuch as theMenin- gitis Research Foundation-Meningococcal Genome Library (MRF- MGL), which includes comprehensive meningococcal genome li- braries for England, Wales and Scotland, have been very suc- cessful in establishing representative epidemiological collections of isolates, which havecontinued to be maintained andupdated (Fig.3a,b).66
PneumococcalAfricanGenomes(PAGe)consortium
Aproject-basedapproachcanalso successfullydelivertraining andcapacitybuilding.ThePneumococcalAfricanGenomesproject (PAGe) is a consortium led by the Malawi-Liverpool-Wellcome Trust, with primary partners inNiger, South Africa, The Gambia, andthe U.K. PAGe focused on genomic analyses ofthe pneumo- coccus across Africa, collecting over 800 isolates, sequenced at theWellcomeSangerInstituteandestablishingcollaborationswith 38countries.InadditiontoencouragingdialoguebetweenAfrican sites,thisprojecttrainedbioinformaticiansincentresinTheGam- bia,MalawiandSouthAfrica.Thisenabledtheanalysisofdataat alocallevel,andrevealedthatserotype1inWestAfricaisdistinct fromserotype 1across the rest ofthe continent.67–71 Furthering epidemiological understanding ofthisserotype, one of the major causesoflife threateningIPDinsub-SaharanAfrica,marksanim- portantcontributiontowardsvaccinetargetdiscoveryanddevelop- ment.
MolecularEpidemiologyforVaccinationPolicy
MolecularEpidemiologyforVaccinationPolicy(MEVacP) isan- other example ofa project-based approachthat aimsto improve globalpublic health by enhancing the diagnosisand surveillance of bacterialmeningitis caused by meningococcus, pneumococcus, H. influenzae and GBS, through building networksin low-income countries, with an initial emphasis on Africa. The aim of this project, funded by National Institute for Health Research (NIHR) andledbytheUniversityofOxford,istoimprovethecharacterisa- tionandvisualisationofoutbreaksacross themeningitisbeltand inform public health vaccination policies, through development
Fig. 3. Meningococcal genomes submitted to the PubMLST Neisseria database, as part of the U.K. Meningococcal Genome Library characterised by core genome MLST (cgMLST), 1605 loci, from 2009 to late 2019 ( n = 4242); (a) coloured by year; (b) coloured by clonal complex.
Table 1
Meningococcal meningitis in 10 African countries, 2018 [Source: WHO weekly bulletin] 76.
Countries # Meningitis suspected cases # CSF performed # Pos # Nm pos %CSF performed % Nm pos % pos
Benin 322 320 7 1 99% 0.3% 2%
Burkina Faso 2421 1590 211 41 66% 3% 13%
Cameroon 1060 111 5 2 10% 2% 5%
CAR 467 699 37 3 150% 0% 5%
Chad 401 304 102 34 76% 11% 34%
Ghana 987 910 89 39 92% 4% 10%
Mali 755 707 126 13 94% 2% 18%
Niger 1496 1151 543 447 77% 39% 47%
Nigeria 4516 804 310 257 18% 32% 39%
Togo 683 1679 38 3 246% 0.2% 2%
All countries (25) 20,843 8650 1531 850 42% 10% 18%
andimplementationofaPubMLST-associatedweb-basedplatform
‘African Meningitis Epidemiology in Real Time’ (AMERT).AMERT will operate a peer-to-peer private website similar to the Euro- peancounterpart,EMERT72,whereby datainthe systemareonly availabletosubmitters.Withrestrictedaccess,referencelaborato- riesarereassuredthatdatawillremainprivate,whichencourages submissionstobemade. Oncepublishedinjournals,thedatacan bemadepubliclyavailable.
Strengtheningcountrylaboratorycapacity
Currently, molecular surveillance inLMICs is largelydriven by expertinternationalcentres based inrichercountries. In general, LMICs frequently lack laboratory capacity, and have limited ac- cessto bioinformatic toolsand support, thus limitingopportuni- tiesforusingsequence-basedapproachesandtheinterpretationof sequencingdatatoidentifythecausativeagentsofdisease.60Exist- ingcapacityneedstodevelopfromarelianceonexternalsupport, towardssurveillance that isdirected and delivered by theregion in the short term and by individual countries in the long term.
Tofacilitatethis, national and regionallaboratories withcapacity toperform molecularsurveillance need tofirst beidentified, and connected,before progresscanbe madeata countrylevel.How- ever,todecentralisecapacitytothecountrylevel,greaterresource andtechnicalsupport isrequired atall levels toachieve sustain- ablemolecularsurveillancebasedonWGS.Therapiddevelopment ofthetechnology andthesupporting software mayassistin this process.
The WHOCollaboratingCenters(CCs), theNorwegian Institute forPublicHealth,theU.S.CDC,theInstitutPasteurParis,theMed- ical Research Council Unit The Gambia at the London School of HygieneandTropical Medicine(MRCG atLSHTM),andtheGlobal ReferenceLaboratory for the IB VPD Surveillance Networkatthe U.S. CDC have laid the foundation for global meningitis genome surveillance by strengthening country capacity for surveillance, laboratory diagnostics and outbreak investigation, particularly in Africa. However, the recovery of isolates from epidemic areas of theAfricanmeningitisbelt remainslow.73,74 Datafromthe WHO weekly bulletins show a low but rising proportion of reported cases with cerebrospinal fluid (CSF) samples obtained and anal- ysed,from∼10%tonearly40%overthepast13years.Whilethere hasbeensomeimprovement,stillinveryfewoftheanalysedCSF samplesisthepathogenidentifiedandisolatepreserved(Fig.4).75 In2018,7.4%ofthecaseshadanidentifiedaetiologycomparedto 2.4% in 2007. Substantial variation among countries is also seen regardingtheproportionofcasesthathavelumbarpuncturesper- formedandpathogensisolated(Table1).76
The transferofclinicalspecimenstonationalandinternational laboratorieswheresequencing canbeperformedisalsochalleng- ing. Typically, the WHO CCs receive isolates from < 10% of re-
portedcases(Table2).75 Duetothechallengesinperforming cul- ture, manylaboratoriesare introducing real-time,or quantitative, PCR (qPCR),bypassingtheneedforculture,whichatpresentisa prerequisite for routine WGS. WGS requires viable isolates to be obtainedwhichcanbeproblematic,particularlyifthepatienthas receivedantimicrobialtherapy.
MenAfriNet,77 an international consortium that supports the strategic implementationof case-based meningitissurveillance in Burkina Faso78,79 Mali, Niger80, Togo, and Chad, also plays an essential role in strengthening global meningitis surveillance. By 2017, MenAfriNet had enroled 33 million people in case-based surveillance81, and 92% of suspected meningitis cases had a CSF specimen collected, of which 26% were laboratory confirmed as N.meningitidis(56%);S.pneumoniae(40%);orH.influenzae(4%).82 This does, however, demonstrate that despite most cases having CSFcollection,threequartersofcasesstillhadanunidentifiedae- tiology.
Initial priorities for capacity building includeimproving labo- ratorycapacityforandstandardisation ofculture andisolationof pathogens,ascurrentlyfewlaboratoriesinsub-Saharan Africacan performculture-baseddiagnosticassays.Onlyafterthis,shouldef- fortsbefocusedonenablingDNAextractionandsequencing.Inad- ditionto resourcerequirements,increasedlaboratory capacityre- lies on technicalsupport from WHO CCsand other international partners, whocan providetraining, qualitycontrol andmolecular characterisation when needed. However, currently WHOCCs lack permanentfundingandasaresultseveralin-countrylaboratories have large and potentially valuable collections of isolates which havenotbeensequenced.Ataminimum,regionalsequencingca- pacitywouldbebeneficialsincesequencingisfundamentaltorep- resentativegenomicsurveillance.
Challengestoachievingrepresentativedatacollections
EnsuringthatWGSrepositoriesareup-to-date andrepresenta- tive is a priority, but likely to be challenging to achieve despite global partners’ support. During outbreaks, rapid sharing of ge- nomic sequence data is crucial fortracking disease transmission and responding to developing health emergencies. In particular, public healthofficials andcliniciansneedto knowwhethercases are likely to be covered by available vaccines as they occur. To achieve this, amuch higherproportionofisolates wouldneed to besequencedtoWGSlevelthaniscurrentlypossible.
Current genome repositories include sequences from collec- tions undertaken for particular public health or research initia- tives, unsystematic collections from laboratories or hospitals, as wellasepidemiologicallyrepresentativecollections.Allcollections are valuable for surveillance of important disease-causing organ- isms; however, systematic andcomprehensive collection and se- quencingfromasmanycountriesaspossibleistheaspiration.
Fig. 4. Meningitis surveillance, 2006–2018: data on suspected meningitis cases from 18 countries south of the Sahara [Source: WHO weekly bulletin] 75. Table 2
Meningococcal isolates sent to WHO-CC’s from the laboratory-confirmed cases reported to WHO [Source: WHO weekly bulletin] 75.
2011 2012 2013 2014 2015 2016
Countries Cases Isolates Cases Isolates Cases Isolates Cases Isolates Cases Isolates Cases Isolates
Mali 29 6 94 30 6 0 12 0 23 16 44 0
Burkina Faso 257 41 843 167 180 20 210 4 258 11 176 5
Niger 373 17 22 0 11 0 24 0 1390 102 352 0
Nigeria 4 0 4 0 10 7 38 5 20 9 22 14
Chad 104 47 47 23 3 24 0 1 1 1 1 0
Cameroon 92 0 19 4 2 0 0 0 6 0 2 0
Benin 0 0 6 41 5 0 4 0 4 0 13 0
Togo 2 0 9 0 4 0 1 16 36 12 307 42
Ivory Coast 0 0 89 7 0 0 0 0 2 1 3 0
Guinea 0 4 0 0 15 9 13 0 74 0 13 0
CAR 0 0 4 0 0 0 0 0 0 7 56 23
Total 861 115 1137 272 236 60 302 26 1814 159 989 84
% isolates 13% 24% 25% 9% 9% 8%
Choiceofclinicalspecimenandassociatedchallenges
Resource-poorsettingsfacemultiplechallengesincontributing isolatestosequencingrepositories.Onesuchchallengeisthediffi- cultyinobtainingviableculturefromCSF,partlyduetorestrictions orreluctanceinperforminglumbarpunctures.
Bloodsamplesofferanalternativesolutiontoincreasethepro- portionofcasesforwhichspecimensareavailable;however,bac- terial load can limit successfulculture recovery due to: (i) early infectionwhen theorganism maynotbe sufficientlyabundantin blood;and(ii)priorantibioticusewhichclearstheorganismmore rapidlyfromthebloodstreamthanfromtheCSF.Inchildren,blood PCRmaybepositiveforpneumococcusasaresultofheavyasymp- tomatic nasopharyngeal colonisation alone.83,84 Additionally, few laboratoriesinlow-incomecountriesarecurrentlyabletoperform blood culture. Dueto such limitations,blood samplingshould be considered asan additionalmeans ofconfirmation, ratherthan a replacementforCSFtesting.
AnothersolutionwouldbetosequencebacteriafromCSFsam- ples directly. Metagenomic approaches were shown to be effec- tive inarecentoutbreakinLiberia, wherenoisolateswere avail- able.11,85Whilecontaminationwithnon-meningitiscausingorgan- ismsorhumanDNAcouldreducesensitivityandraisepatientpri- vacyissues,targetedsequencing approachesorremovalofhuman sequencedataaftersequencingwouldmitigatethisrisk.
Visionforaglobalmeningitisgenomepartnership PublicsharingofWGSdata
A global representation of sequence-based typing including WGSis crucial; however,unique isolate identifiersshould be de- finedtoallowisolatestobetrackedinternationallywithingenome repositories.Thisisessentialfortheeffectivecontrolofepidemics, vaccine evaluation and development. A global overview also re- quires genomesequence datato be open access, assembled ina readilyinterpretable andanalysable manner, andassociated with appropriate metadata.1 Data holders are also expected to upload shortreaddatatotheENAsothatdataareopentoscrutiny.While the data privacy issue is complex,the requirementfor metadata hastobalanceprivacyconcerns,withtheobjectiveofenablingthe interpretationofgenomicdataforpublichealthbenefit.Anagree-
1Metadata: provide an in-depth controlled description of the sample that your sequence was taken from. Essentially the ‘what, where, how and when’ of your study from collection to sequence generation, plus contex- tual data such as environmental conditions or clinical observations. ( Definition from the European Bioinformatics Institute, Available at: https://www.ebi.ac.uk/
training/online/course/ebi- metagenomics- portal- submitting- metagenomics- da/
what- are- metadata- and- why- are- they- so- im- 0 )
mentontheminimumcoresetofassociatedmetadatathatshould bedepositedwithWGSdataisneeded.
Clearcodesofpractice(COP)shouldbeestablishedforusersof repositorieswithin the partnershipand fordata publication. The consensusCOPguidelinesattheirbasiclevel,servetosecurepub- licaccessby defining thepermitted usesof repositories,thus re- assuringsubmitters who mayhavelegitimateconcerns aboutthe useoftheirdata.
Curationandstratificationtoensuredataquality
The active curation of genome libraries is essential to ensure dataqualityandenableuserstointerpretsequencedatameaning- fully.While automationcan reduce theburdenof curation,man- ualcurationwillberequiredforasmallpercentageofdataforthe foreseeablefuture.Stratification,throughclearlabellingofgenome librariesisessential toensure that usersunderstandwhat theli- brariescontain,howthey arearrangedandtheoriginsoftheiso- latesthey arestudying,i.e.carriageordiseaseisolates, systemati- callycollectedornot.
Overlapamongdifferentgenomicplatformsanddatabases:the importanceofdatalinkage
Existinggenomicdatabasesoverlapincontentandhavediffer- ent but partly-overlapping functions. The interpretation of these databasesdoes,however,relyontheongoinggenerationandmain- tenance of a unified nomenclature. For example, a fundamental function of PubMLST is the ability to allow comparisons to be madeamongisolateswithinaspeciesandfrommultipledatasets.
Datalinkageamongexistingdatabases,facilitatedbyAPIs,ises- sentialtoachievedataharmonisation,avoidfragmentationofinfor- mation,andenableinterpretation.However,whileAPIshavemany advantages,acaveattotheiruseisthatsporadicdownloadofdata maymeanthatusersareunawareofthedatasource,andunknow- inglyconductanalyses onun-representative samples,riskingmis- leadingresultsandunwarrantedassumptions.
Aspirationforasingleinterface
For a GMGPto attract submissions on a globalscale, there is an aspiration for a single interface that successfully draws from publicly-availablegenomicandprovenancedata; however,portals to facilitate this are needed. Combined with visualisations, such asystem wouldalso make the valuable informationgained from WGS data more widely accessible, e.g. to public health officials, ministriesofhealth,globalhealthexperts,science/health journal- istsandthebroaderpublic.Multiplelevelsofcomplexityare,how- ever,required fordifferent users, such that experts can continue toundertakeadvanceddataanalysis,whilenon-expertgroupscan engagewithrelevantpublichealthinformation.
Conclusionsandnextstepsinthedevelopmentofaglobal meningitisgenomepartnership
WhiletherehasbeenariseintheuseofWGS,thishasmostly beendriven by a smallgroup ofcountries withadequate capac- ityandresourcestoparticipateandbenefitfromperformingWGS.
Asa result,currentWGScollections arenot representativeofthe globalmeningitis picture. The advent of the Global Roadmap to Defeat Meningitis provides an excellent opportunity to create a newvision for the role that sequence-based approaches, includ- ing WGS,can play in helping to defeat meningitis by 2030, par- ticularly through improvingthe globalsurveillance of meningitis.
Asteering group is beingestablished to coordinatethe initiative
and encourage high-quality data curation. Next steps in the de- velopmentofapartnershipwillinclude:(i)developingguidelines on open-access sharing of genomic data, (ii) defining a core set ofmetadata to facilitateopen data sharingwhile protectingcon- fidentiality(iii)engagingwithcountriesnot currentlyinvolvedin existing repositories; (iv)improvingcountries’ capacityto culture andisolatepathogens;(v)considerationoftheneedforadditional WHOCCsindifferentregions;(vi)improvementsinvisualisations to provide data in a way people can understand while engaging publichealthbodies,governments,epidemiologists,scientistsfrom other fields,journalists,andthepublic;and(vii)establishmentof apooledfundtoenablecapacitybuildingandcountry-drivenen- gagementwiththepartnership.
DeclarationofCompetingInterest
AMisanemployeeoftheGSKgroupofcompanies.
AvdEhasreceivedgrants fromPfizer, consultancyfeespaiddi- rectlytotheinstitutionfromGSK andparticipatedinScienceAd- visoryBoardsforPfizer,GSKandSanofi Pasteur.
ER, LG & VS representMeningitisResearch Foundation, which receivesgrantsfromSanofi Pasteur,GSKandPfizer.
JF is an employee of Pfizer Incand mayhold stock/stock op- tions.
JL & RB perform contract research on behalf of PublicHealth EnglandforGSK,PfizerandSanofi Pasteur.
JV acts as temporal advisor and receives grants for research fromSanofi-Pasteur,NovartisVaccines,GlaxoSmithKlineandPfizer, payedtohisinstitution.
LHHhasservedasaconsultanttoGSK,Merck,Pfizer,andSanofi Pasteur.
LSis currentlyemployed by theGSK group of companiesand mayholdGSKsharesaspartofheremployeeremuneration.
POisanemployeeofSanofi Pasteur.
SDB,HBB,SB,ABB,DAC,LF,OBH,RSH,MJvR,KAJ,BKA,SL,MLF, ML,NM,JM,MCJM,LMS,CMCR,OR,XW,SYandJMShavenocon- flictsofinterest.
Acknowledgements
Thefindingsandconclusionsinthisreportarethoseoftheau- thor(s)anddonotnecessarilyrepresenttheofficialpositionofthe CentersforDiseaseControlandPrevention.
Funding
This research did not receive any specific grant from funding agenciesinthepublic,commercial,ornot-for-profitsectors.
References
1. Zunt JR , Kassebaum NJ , Blake N , Glennie L , Wright C , Nichols E , et al. Global, regional, and national burden of meningitis, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 2018; 17 :1061–82 . 2. van de Beek D . Progress and challenges in bacterial meningitis. Lancet
2012; 380 :1623–4 .
3. Global Burden of Disease Collaborative Network. Global burden of disease study 2017 (GBD 2017) results , Seattle: Unites States Institute for Health Metrics and Evaluation; 2019. Available from http:// ghdx.healthdata.org/ gbd- results- tool [Last Accessed: February 2020] .
4. Mwenda JM , Soda E , Weldegebriel G , Katsande R , Biey JN , Traore T , et al. Pedi- atric bacterial meningitis surveillance in the world health organization african region using the invasive bacterial vaccine-preventable disease surveillance net- work, 2011–2016. Clin Infect Dis 2019; 69 (Supplement_2):S49–57 .
5. Trotter CL , Lingani C , Fernandez K , Cooper LV , Bita A , Tevi-Benissan C , et al. Im- pact of MenAfriVac in nine countries of the African meningitis belt, 2010–15:
an analysis of surveillance data. Lancet Infect Dis 2017; 17 :867–72 .
6. Funk A , Uadiale K , Kamau C , Caugant DA , Ango U , Greig J . Sequential outbreaks due to a new strain of Neisseria meningitidis serogroup C in northern Nigeria, 2013-14. PLoS Curr 2014; 6 .
