TRanslational bioinformatics with Artificial INtelligence – TRAIN

Research Activity

The group’s activities focus on the processing and integration of complex biological data using advanced computational models. These models are built using the most modern Artificial Intelligence (AI) approaches, which include machine learning, multivariate statistical analysis, predictive modeling, and the latest deep learning architectures, such as graph networks and attention-based models. Particular emphasis is placed on introducing Explainable Artificial Intelligence (XAI) approaches, which are essential for ensuring transparency, interpretability, and clinical trust in the use of algorithms, especially in high-criticality decision-making contexts such as medicine.

From this perspective, the research addresses complex challenges such as: the semantic and biologically significant representation of features, the choice of the most appropriate learning strategies (supervised, unsupervised, semi-supervised, transfer learning) along with the definition of the most suitable architectural models, and the explicit statement of the inferential logics underlying the predictive models, in order to make computational decisions interpretable by clinicians and biologists.

A transversal axis of the research is represented by the use of network biology as a paradigm for the integration and holistic study of multimodal data, with the aim of explicitly modeling the relationships between heterogeneous biological components (genomic, transcriptomic, proteomic, phenotypic, etc.) and supporting the interpretation of results through structured representations that are both interpretable and adherent to the complexity of biological systems.

In accordance with the principles of Open Science, the research group is committed, where possible, to making the data and methods used and developed freely accessible for transparent and reproducible sharing with the scientific community. The prototypes developed will therefore be preferably released in the form of accessible tools or web services, promoting their use within the international scientific community and facilitating the adoption of advanced bioinformatics solutions in the clinical-translational context.

Goals

In recent years, the interdisciplinary research areas of bioinformatics and computational biology have assumed a central role in the medical landscape, particularly within the field of precision medicine. Here, translational bioinformatics acts as a bridge between biological research and clinical application, aiming to transform complex biological data into diagnostic, prognostic, and therapeutic tools useful in medical practice.

In this context, the multidisciplinary expertise of the research group members—both in the biological domain and the computer science/computational domain—coupled with consolidated experience in computational biology over the years, allows the group to pursue its main objectives.

Specifically, the general objective concerns the development of strategies, tools, and methodologies capable of generating new knowledge useful for understanding the molecular mechanisms of complex diseases, as well as supporting the development of personalized therapeutic solutions—the so-called targeted therapies—through the management, integration, processing, and interpretation of clinical and multi-omics data (e.g., genomic, transcriptomic, metabolomic, etc.).

Application Fields

Precision Medicine, Computational Biology

Back To Top