Roadmap for development of disease control strategies for bTB

What are we trying to achieve and why? What is the problem we are trying to solve?

• Use genomic information of field isolates to construct phylogenies and answer evolutionary and epidemiological
  questions.
• Unravel transmission dynamics using phylogenetic approaches:tracing of potential sources for outbreaks and
  determining if an outbreak is ‘local’ or comes from a more distant source
• Alongside these epidemiological uses for genomic data, there is also potential to investigate evolutionary questions. Sequence data can help to infer past population demographics – bottlenecks caused by test and slaughter schemes, expansion of certain lineages etc. Data can also be used to search for signals of selection in the pathogen genomes that might underpin novel phenotypes of epidemiological consequence. The latter is speculative but possible owing to similar findings with M. tuberculosis. It is recognised however that searching for signatures of selection in such genetically homogeneous, slowly evolving pathogens as the MTBC is difficult.

What are the scientific and technological challenges (knowledge gaps needing to be addressed)?

1. Collection of representative genomic datasets from multiple affected livestock and wildlife hosts at various geographic scales is necessary to facilitate the transmission dynamics and general disease source tracing applications. This underway in many jurisdictions, but not all. Funding not the same in all countries.
2. Standardised nomenclature across regions for M. bovis as has been achieved for M. tuberculosis would facilitate greater utility of datashare between countries.
3. Standardising wet lab and bioinformatic approaches to ensure comparability between regions would be helpful. There is likely no need to all use the exact same software etc, but some degree of benchmarking through ring trials etc would at least help to ascertain that different labs are producing comparable data.
4. The low mutation rate of M. bovis can make inference of transmission dynamics and disease source tracing difficult – more appreciation of limitations needs to be to the fore. Genomic data needs to be paired with other rich meta-data on animal movement etc to help inform policy makers and vets on the ground.
5. Reaching a consensus on which methods to use to robustly construct home ranges for phylogeographically localised lineages will aid epidemiological tracing applications.
6. Similar to above, reaching consensus on SNP cutoffs necessary to define lineages and outbreak clusters should be encouraged.

What approaches could/should be taken to address the research question?

1. Efforts to standardise nomenclature should be undertaken across multiple territories. Some efforts in Europe already underway.
2. Setting up of ring trials for blind sequence comparison to benchmark laboratory and bioinformatic processes.
3. Mobilise collated data for analyses beyond just epidemiology.

Train teams of people to undertake analyses seeking signatures of selection in M. bovis genomic data. If candidate loci are identified, design experiments to assess if they have epidemiologically significant phenotypic outcomes. Knowing the latter might have utility in improving eradication schemes.

What else needs to be done before we can solve this need?

Having geographically wide sampling of common M. bovis lineages – temporal depth is not such a big concern however.

Existing knowledge including successes and failures

Various updates of the AF2122/97 reference M. bovis genome have been completed over the last decade. This reference sequence provides the template against which SNPs are called from raw genome sequence runs of field isolates, enabling the construction of phylogenies which can be used to answer
evolutionary and epidemiological questions.
Several studies have used genomic data from outbreaks to inform on transmission dynamics in multi-host systems.
Principally the UK and Ireland and French epi-systems involving cattle, badgers, wild boar and deer have been investigated.
These phylodynamic approaches are helping to determine which disease transmission routes predominate. They are dependent however on having well sampled epi-systems from multiple hosts over a long period of time. The latter point is crucial to facilitate detection of a temporal signal with which time stamped phylogenies can be produced.
Alongside transmission dynamics, there are more ‘everyday’ uses for M. bovis genomic data – principally in tracing of potential sources for outbreaks and determining if an outbreak is ‘local’ or comes from a more distant source. The striking phylogeography of M. bovis aids this type of work. It is dependent however on having geographically wide sampling of common M. bovis lineages – temporal depth is not such a big concern however. It is recognised however that searching for signatures of selection in such genetically homogeneous, slowly evolving pathogens as the
MTBC is difficult.