Machine-learning guided immune fingerprinting for rapid detection of life-threatening infection in liver cirrhosis

HYPOTHESIS

Early prediction of liver disease through machine learning techniques applied to routinely collated data will improve outcomes and enable improved screening. In advanced disease similar machine learning techniques will enable rapid diagnosis of life-threatening bacterial infections in cirrhosis

AIMS

- Develop biomarkers of bacterial infections in advanced cirrhosis using serum cytokines to personalise therapeutic approaches using classification methods
- Develop an in-depth analysis and key factors that predict the progression of liver disease using machine learning techniques
- Develop epidemiological database of all liver diseases and investigate progress of them

4. OBJECTIVES
- Develop a dataset of peripheral blood and ascites lymphocytes and monocytic cells and cytokines in advanced cirrhosis
- Develop machine learning protocol to apply to this dataset and interrogate it for key biomarkers that can be applied in clinical practice.
- Investigate the natural history of liver disease in Wales from routinely collected clinical data
- Machine learning prediction of high-risk liver diseases progression

EXPECTED OUTCOMES

The anticipated results contain flow cytometry events and a biological meaningful prediction of finger-prints of infection.