The main purpose of my PhD is the development of computational tools for bioinformatics analysis of clinical and omic data for the prediction of genotypic associations/phenotypic and prognostic, using machine learning techniques. I have been involved in projects in the cardio-metabolic area, in particular I have worked on methylation data and phenomenological variables to predict myocardial infarctions up to 15 years in advance. I have also worked on predicting the gravity of liver fibrosis using only metabolomic variants and markers of fibrogenesis. I carried out more technical studies: one on the upper limits of the metrics used to evaluate the goodness of the predictions at real values and one on the calibration of the variant-scoring predictors.