Final Report Summary - BIONET (Network Topology Complements Genome as a Source of Biological Information)
Large-scale molecular and medical data sets are increasingly becoming available. The question is what can we learn from these complex data about biology and our health? The project developed new methods for analytics and integration (fusion) of interconnected biomedical data represented as networks. The new methods were applied to address current problems in computational biology and medicine, including molecular network comparisons and alignments, which are key to comparing and understanding of our cells. The newly developed network science and machine learning algorithms correlated the wiring patterns of proteins in molecular networks with their biological function and involvement in diseases, yielding biological insights. Also, the new methods for network data integration uncovered new relationships between diseases based on modern, systems-level molecular data and improved stratification of patients of the same cancer type into clinically significantly different subgroups, biomarker discovery, and re-purposing of the existing drugs to different patient groups based on the evidence newly discovered from heterogeneous molecular and clinical data collectively, addressing some of the foremost problems in precision medicine. The project demonstrated complementarity of different types of molecular data in uncovering new biomedical knowledge, including the complementarity of genetic sequences and the wiring patterns of molecular networks. It opened up important new research questions stemming from this complementarity of the data coupled with the complexity of the computational problems for which solutions still need to be designed to collectively mine the data.