Sara C. Madeira

Complex diseases, such as neurodegenerative diseases, present unique challenges to medicine due to their devastating effect on patients and their families, and socio-economic impact in modern societies. As such, identifying disease risk profiles together with understanding the individualized disease diagnostic and prognostic patterns, is top priority on ongoing research. In this context, high throughput technologies introduced a major change in biology and medical research paradigms. Systems biology has been successfully studying omics, aimed at a fundamental understanding of biological processes and modeling of biological networks. However, it has focused primarily on the molecular scale disregarding the effects of physiology and its enormous variation among individuals, produced by the interaction of each person’s biology and experiences through life. Systems medicine, on the other hand, has recently emerged as the application of systems biology to human health, by incorporating clinical information into its models. Moreover, several authors emphasised the need for integrative approaches, able to infer relationships between omics, clinical, and personal data, to provide a broader understanding of complex diseases. Promising attempts integration wise are arising but no effective integrative approaches between clinical and omics data exist.

This project plans to further study and develop methods for prognostic prediction in complex diseases based on biomedical data analysis (omics and clinical), where data are characterized by a set of heterogeneous biomarkers collected for patients during their follow-up. This is the type of data made available by international consortiuns, such as Alzheimer’s Disease Neuroimaging Initiative (ADNI) and Pooled Resource Open-Access ALS Clinical Trials Database (PRO-ACT), and analysed in NEUROCLINOMICS, a grant funded project I coordinate in Portugal, with a focus on clinical data and neurodegenerative diseases. The motivation for this proposal is the fact that, the efficient analysis, from a temporal perspective, of this type of heterogeneous and multidimensional data, composed by a set of snapshots collected periodically for a set of patients (follow-up), where each snapshot corresponds to the analysis of an heterogeneous set of biomarkers, has been little explored, besides its recognized importance to the discovery of patients’ profiles and progression patterns that can be effectively used to develop automatic prognostic methods.