ETD

• Graph learning
• Machine Learning
• Molecular fingerprint
• Schizophrenia

The thesis discusses the use of Machine Learning models to predict compounds that can reduce glial phagocytic activity in Schizophrenia patients. The used approach is based on the generation of a molecular fingerprint through a graph Machine Learning technique called graph convolution. The analysis has been performed in collaboration with the SPARK research group at Stanford, that has provided the data.