We develop methodological tools, algorithms and software to analyze data and vital signals reflecting individual health and to exploit them with learning techniques to improve patient therapy, possibly creating real-time feedback mechanisms. Tools are personalized, adaptive, proactive and equipped with intelligent self-diagnostic functions. Models to predict and prevent the incidence of new diseases or medical complications are also investigated.
Deterministic and stochastic modelling of transcriptional networks and signaling pathways, reverse engineering of transcriptional networks and miRNA-mRNA regulatory motifs, and integration of genetic phenotypic and environmental risk factors via Systems Genetics approaches
Advanced data mining and machine learning methods for robust biomarker discovery, predictive modeling and clustering from microarrays and next generation sequencing data
This research area deals with the use of mathematical modeling techniques to develop and validate mathematical models able to either simulate the behavior of complex, dynamical biological systems, or to estimate key parameters usable to quantify physiological processes. Depending on the aim and the available experimental data, the models can be whole-body, organ/tissues, cellular or multiscale, formulated as ordinary or partial differential equations, deterministic or stochastics, minimal or large scale.
Development of control algorithms for automatic insulin delivery in type 1 diabetes (Artificial Pancreas): personalization, run2run adaptation, fault detection and fault compensation for safe unsupervised AP use.
In silico validation and clinical testing of AP algorithms.
