A systems biology approach to reveal cellular pathways underlying Alzheimer disease

Executive Summary of Scientific Report A systems biology approach to reveal cellular pathways underlying Alzheimer disease BSF-2011296 Esti Yeger-Lotem, Ben-Gurion University of the Negev Susan Lindquist, Whitehead Institute for Biomedical Research and MIT Alzheimer disease (AD) is the most common neurodegenerative disorder affecting 35 million people worldwide. In an effort to elucidate the cellular pathways underlying AD, genomic and transcriptomic profiles were collected from patients and compared to healthy individuals. Since these types of profiles were often analyzed in separation, important inter-connections might have remained hidden. In this grant, we developed and applied network biology tools to reveal the pathways underlying AD and other tissue-selective hereditary diseases. Our studies resulted in the following publications and findings: a. We developed a context-sensitive framework for the analysis of human signaling pathways in molecular interaction networks (Lan et al., Bioinformatics 2013). This framework was unique using the different gene ontology annotations of proteins in a pathway to identify and rank context-sensitive paths in the interactome, thereby revealing intermediate proteins and pointing to their relevant context.

b. We developed the MyProteinNet web-server, which allows users to create context-sensitive protein-protein interaction networks (interactomes) for human, yeast, and multiple other organisms (Basha et al., Nucleic Acids Research 2015).

c. We developed a scheme for differential network analysis, applied it to data of human tissue interactomes, and showed that this scheme helps prioritize genes involved in tissue-selective hereditary diseases (Basha, Argov et al., in preparation).

d. We applied our integrative network-biology tool ResponseNet to find genes that connect known AD-associated genes from OMIM database to AD-associated genes from AlzGene (based on genome-wide association studies, containing only one common gene) in a brain-specific differential interactome. ResponseNet predicted 13 intermediate nodes, for which we found evidence for 12 genes being associated with AD.

d. Upon analyzing 112 tissue-selective hereditary diseases, we found that dosage relationships between causal genes and their paralogs may affect their phenotypic outcome, pointing to paralogs as potential modifiers of disease manifestation (Barshir et al, 2017, bioRxiv).

e. Lastly, we applied ResponseNet to reveal the activities of the yeast kinase Rio1 (Iacovella, Bremang, Basha et al., in revision).