Year 2 progress highlights:

• Analysing large-scale biological data (WS1):

  • New methods for combining results from multiple “omics” studies (such as genomics) have been developed. These methods account for differences between studies, improving reliability – something past approaches could not do. A draft paper is nearly ready for submission.

  • A scalable k-mer workflow has been built, automating the process of retrieving bacterial genome data, assembling genomes, and comparing classification methods. This helps scientists assess the accuracy of tools used to group bacteria by lineage.

  • A genome re-annotation tool has been modularised to support multiple annotation strategies and improved with a draft of GFF3 format support, making it more versatile for different research needs.

• Infectious disease modelling (WS2.1ffi):

  • A model of infection surveillance has been refined so that it uses information typically collected in epidemiology, making it more practical for real-world use. Improvements also fixed rare cases where the model could fail. The model is now being tested for accuracy.

• Sensor data and real-time monitoring (WS2.2):

  • Work continued on analysing sensor data from livestock (using the Langhill herd dataset), with a review paper on quantitative methods for real-time monitoring in agriculture due for submission in March.

• Behavioural modelling (WS2.3):

  • A new framework, behave*, was created to model behaviour more effectively. It includes six key categories: behaviours, environment, history, algorithm, values, and entity-state. This framework is being tested in two areas: modelling how livestock behaviour drives the spread of infectious disease, and understanding how land management decisions are made.

This project is advancing the development of genomic, statistical, and modelling tools to better understand infectious diseases, antimicrobial resistance, and environmental monitoring. The work spans literature reviews, new software development, and the creation of models with real-world applications in health, agriculture, and ecology.

Year 1 progress highlights:

• Genomics and software tools (WS1):

  • A review of nearly 500 studies on the use of public genomic data identified 46 key papers, helping map how these data are applied across research fields.

  • A new approach to k-mer analysis (a key method for analysing genetic data) has been developed as part of a software suite called KPop, now under peer review.

  • A first version of a genome reannotation tool has been completed, already in use at the UK Health Security Agency to reanalyse pathogens such as monkeypox and flu. The tool will be expanded to handle more complex viral genomes.

• Infectious disease modelling (WS2.1):

  • New methods have been developed to improve the design of challenge studies, which are used to estimate how genetic factors affect disease spread.

  • A model to estimate whether a farm, region, or country is truly “free from infection” has been refined, using an efficient simulation approach (Fleming-Viot method).

  • A mathematical model of antimicrobial resistance (AMR) has been built, showing how resistant and non-resistant bacteria can co-exist – a crucial step for realistic modelling of AMR dynamics.

• Sensors and behaviour (WS2.2 & WS2.3):

  • A review of almost 250 studies on sensors found only a few deal with real-time prediction, highlighting opportunities for future development in agriculture, health, and environmental monitoring.

  • Work on modelling behaviour at both individual and system levels revealed approaches ranging from simple parameter changes to complex “emergent” responses. Several gaps in this area have been identified as priorities for further research.