1. Global, real-time microbiological genomic identification and control project
Frank M. Aarestrup
Center for genomic Epidemiology 1 1

2. Bacterial characterisation is used for
Predict treatment possibilities
Continuous surveillance
Predict virulence
Tracing routes of transmission
Control of cause of treatment failure
Analyse population structures

3. Bacterial identification and typing ... Evolution of typing methods
IS-typing
flj-typing
RFLP
VTNR
REA
PFGE
Plasmid
profiles
MLEE
RAPD
Phage
typing
MLST
AFLP
Bacteriocin
typing
Serotyping
1920
1940
1960
1970
1980
2000
Evolution of typing methods

5. 5

7. World Largest HUS Epidemic Due to EHEC
Germany (RKI, July, 26th, 2011)
EHEC
3,481 cases
18 deaths
HUS
852 cases
32 deaths
Europe / North America (WHO, July 21st, 2011)
89 cases
no deaths
52 cases
2 deaths
Cases
Deaths
EHEC
HUS
RKI, July 26th, 2011, adapted [

8. Typing strategies Cheap methods for surveillance
Outbreak investigated by more than one method
No common method combination applies to all situations

9. E. coli, Salmonella
Campylobacter
Neisseria
Different clonal populations

10. Purpose of Center for Genomic Epidemiology
Provide a proof of concept of combining bioinformatics with global epidemiology in real-time
And provide a useful facility for frontline users

12. Store in database
Gb upload
MLST
Novel typing
Phylogeny
Etc
De novo assembly
Template bases assembly
Resistance
Unique project ID (= Directory)
16 S typing
Client side assembly/
compression
Novel pathogenecity
Geo mapping ++
Virulence
Real time return of preliminary data (JS tasting on files)

13. List of genes (100% or >95%) Theoretical resistance phenotype
Resfinder
NGS
Illumina
PacBio
454..
Assembly pipeline
Resistance gene profile
Sanger
Fasta
Fasta
List of genes (100% or >95%)
Accession numbers
Theoretical resistance phenotype
Optional BLAST output
Sanger

14. Front page

15. Results 16

16. Computing needs Denmark (1 million sequences/year):
2h CPU time + 10mb storage for 1,000,000 genomes per year in 6 years = 230 Cores, 60TB storage needed
EU (100 million sequences/year)
2,300 cores, 6PB storage
Global needs (1 billion sequences/year)
23,000 cores, 60PB storage ~ 30th biggest computer (Smaller than Airbus’)

17. Building global partnership
Global consensus meetings
Bruxelles (Sep. 2011)
Washington DC (March 2012)
Involving partners as curators of in silico arrays and/or specific species
Proof of concept applications and projects
Case stories
Hospital diagnostics
V. cholerae from Nepal
MRSA CC398 phylo-evolution
Evalution on Campylobacter, E. coli, Salmonella

18. Washington – March 2012
>100 participants covering, governmental institutions, universities, hospitals and companies
FDA, CDC, FBI, NCBI, DoD, USDA, PHAC,
Harvard, Maryland, Virginia, Los Alamos, TGen, Aligent, Broad
EBI, eCDC, Sanger, Oxford, DTU, RIVM, FZB
BGI, DDBJ, etc

19. 20 November 201820 November 201820 November 2018

20. Conclusions There is a need and we believe it is possible
Need to broaden discussion to a larger forum
Need for special groups working on
Global participation (198 countries!!!)
Repository (NCBI, EBI, DDBJ)
Access (Political, legal)
Output (simple & advanced, diagnostic & epidemiology, genes & species)
Taxonomy

21. When will WGS replace all other techniques?
The - £ $ € - question
WGS – today 100 € going to 50 or 10?
Traditional:
Identification – 10 €
Serotyping – 25 €
Susceptibility testing – €
PFGE – 50 €
MLST – 250 €
Molecular characterisation – 10 – 1000 €
Already competitive
20 November November November 2018

23. 22 publications