Biological data intelligence

HERALD / Johne’s disease research

We analyse biological measurements over time to investigate patterns and potential diagnostic indicators in Johne’s disease research.

Project context

The project and 42able’s role

HERALD (Health Enhancement through Rapid Assessment of Livestock Diseases) is an Innovate UK-supported collaboration with Diagnostig. It investigates data associated with a novel diagnostic approach to Johne’s disease in cattle.

42able contributes software, data analysis and modelling. We examine biological measurements across animals and over time to explore possible disease-related indicators.

Area
Research
Category
Biological data intelligence
Current status
Active collaborative R&D
Stage
Applied research · under evaluation
Back to research

Data under analysis

Biological measurements and sample data

The analysis brings together the available measurements and their context while retaining source values, missing data and uncertainty.

Assay measurements
Biological measurements associated with the diagnostic approach, including the immune signatures and biomarkers under investigation.
Comparator measurements
Data from established comparator methods, used to examine relationships with the measurements being investigated.
Repeated observations
Measurements associated with individual animals and sampling times, supporting analysis of changes within animals and differences between animals.
Sample and animal context
Available contextual records used to interpret observations and examine possible relationships with disease progression.

Analytical work

Statistical and machine-learning methods

We combine data preparation with statistical and machine-learning investigation.

Data structure and quality

Organise measurements, retain their provenance and examine completeness, comparability and the effects of missing observations.

Statistical modelling

Investigate variation and relationships between measurements, taking account of repeated observations and the available sampling context.

Temporal patterns

Examine changes within animals over time and whether their timing may be related to disease progression.

Machine learning and predictive modelling

Explore candidate relationships and potential diagnostic indicators.

Current status

Research status and evidence

The collaboration is active. 42able’s contribution includes preparing and analysing the available data and developing the evaluation approach.

Statistical and machine-learning investigation is ongoing. Potential diagnostic indicators and predictive relationships remain unvalidated hypotheses. Predictive performance, usefulness and diagnostic relevance need evaluation and independent validation.

The published collaboration announcement describes research and evaluation in progress. Completed analysis results, validated diagnostic findings and predictive performance have not been published here.

Research questions

What needs validation?

  • Do changes within an animal add information beyond an individual measurement?
  • Which candidate patterns remain consistent across animals, sampling conditions and stages of infection?
  • Can a potential indicator be distinguished from measurement effects, sampling differences or missing data?
  • What independent data and validation would be needed before a result could inform diagnostic use?

Work with us

Discuss a research collaboration

Contact us to discuss how our data analysis, machine learning or computer vision expertise could contribute to your research project.

Contact 42able