Polygenic and Omnigenic Phenotype Risk Prediction with Machine Learning
Principal Investigator: Miss Chloe Hequet
Approved Research ID: 52480
Approval date: September 18th 2019
A system that was able to predict disease risk from the genetic data of a patient would have many valuable applications in medicine, as it would allow for the prevention and early detection of health conditions. At the moment the effectiveness of disease risk prediction methods is very limited; this may be because these methods do not use enough of the available genetic data to make predictions and therefore miss vital information. Machine learning techniques are well suited to problems involving large amounts of data, and could be incorporated into the process of disease risk prediction to allow more of the genetic information to be used, potentially improving predictions. If a machine learning model was able to predict disease risk better than current methods, analysis of this model could also yield new insights about the genetic basis of disease, informing future research in this field. This work aims to use machine learning techniques to improve the prediction of obesity risk from genetic data.