Towards an autonomous in silico researcher: Using logic modelling to automate the explanation of unexpected results

Biological experiments often produce unexpected results, i.e. results that contradict the researcher's existing understanding. Such results can be revolutionary, revealing new knowledge or leading to new testable hypotheses, but they can also be due to gaps in the researcher's prior knowledge, or experimental error. Explaining unexpected results is therefore critically important. However, today's massive datasets contain so many unexplained results only a minority can be investigated manually, potentially leaving new discoveries unnoticed. This project aims to develop methods that combine experimental data, metadata and prior knowledge to automatically explain all unexpected datapoints within a large dataset, hastening biological discoveries.