Using machine learning technologies to understand mechanisms in fatty liver disease

The aim of the project is to exploit the unique big data opportunities of mass cytometry data to explore the mechanisms behind Non-alcoholic fatty liver disease (NAFLD). This exciting new technology can measure cell markers in over 200,000 cells in each sample and leads to many common “big data” problems. To tackle this data we seek to exploit algorithms and methods developed in other areas of science and engineering, and in particular use current deep learning methods, to identify common patterns of markers within cells across healthy samples and those with NAFLD. We will additionally look to combine this analysis with other datasets such as RNA-Seq to more fully determine the key processes involved. The project is a purely computational project and will involve developing tools using common programming languages such as Python, R and C/C++.

The results from the analysis of this data will guide our understanding in the key mechanisms behind NAFLD. NAFLD is the most common cause of chronic liver disease in the West. The challenge is to distinguish patients who are unlikely to develop significant liver damage from those who will develop the aggressive form, which can lead to cirrhosis and liver cancer. However, there are no tests that can make these predications at early stages and there are no medicines currently available to treat the condition. Although not fully understood, aggressive disease probably develops because of a combination of our environment, lifestyle and genes.