Process & Systems Modelling

A modelling framework to examine traceability in the Scottish sheep population

Following the UK foot-and-mouth disease outbreak in 2001, EU regulations led to electronic tagging of sheep so that individuals can be identified. In Scotland, electronic readers identify individual sheep as they move in batches through critical control points (mostly markets and abattoirs). The readings are then uploaded to a central database (ScotEID).

Operational aspects of electronic identification (EID) inevitably lead to uncertainty, including read errors. However, the impact of these uncertainties on traceability, defined in terms of the ability to determine possible infectious contacts of an identified sheep in the event of a disease outbreak, is far from clear. BioSS has developed a novel, simulation-based, stochastic modelling framework, which enables assessment of traceability in the presence of EID data and other relevant sources of information. Traceability is then interpreted as an emergent property of the modelled system, potentially depending on sheep demography, batch selection processes at holdings, properties of the between-holding network of sheep movements, patterns of infectious contacts between sheep and variations in read error rates.

This formulation enables both an assessment of the current system and the targeting of disease control efforts, while taking into account system uncertainty. Outcomes from various disease control strategies can be compared.

batch movement map Simulated results based on 2011/2012 batch movement data. Estimated levels of contact (darker red means more contact) between a specific sheep residing in the south of Scotland and sheep in all Scottish parishes, given perfect knowledge of individual sheep movements (left) or knowing only electronic readings (right). When only electronic reads are available, the complexity of the sheep movement network and the amount of read uncertainties can lead to many more parishes with non-zero estimated contact. However, parishes at greatest risk can still be identified due to their higher contact estimates.

Further details from: Stephen Catterall and Sarah Brocklehurst

Article date 2015


Statistical Genomics and Bioinformatics

Process and Systems Modelling

Statistical Methodology

PhD Opportunities

Meetings & Seminars