DTEN: a network-based workflow to identify drug target networks from gene expression data in rare disease
Year: 2019
Type: other
Abstract: Identification of drug targets from preclinical ‘omics data remains a chronic challenge within the research community. This challenge stems from multiple causes including inherent noise in genomic assays, inability to generalize experiment-specific data to a broader context, and difficulty in determining specific relationships between gene expression changes and molecular pathway dysregulation. Network- based algorithms can be used to address these points by first integrating ‘omics data with published molecular interaction data and then selecting specific pathways that are relevant to the disease of interest. Integration of protein-protein, protein-DNA or protein-RNA interactions together with network refinement approaches such as those based on network flow, belief propagation, or other graph-theoretical constructs such as Steiner Trees, provide distinct interpretations of molecular data that can enable scientists to develop better testable hypotheses. In this work, we extend existing network tools to incorporate drug-target interactions into network-optimization algorithms that can identify specific drugs of interest from a set of differentially expressed genes. We apply this algorithm to study Neurofibromatosis Type 1 (NF1), a tumor-causing syndrome affecting 1 in 3000 people. In rare diseases such as NF1, there exist modest numbers of samples, giving rise to a greater need for integrative algorithms. Despite the collection of transcriptomic data across numerous types of NF1 tumors, to date, no single analysis has compared all tumor samples to each other. As such, we built a scientific workflow comprised of existing network tools and customizations to compare these data and identify specific drugs that putatively target these diseases.
Institution: Sage Bionetworks
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W2968092997