Data Management and Bioinformatics Core

点击翻译按钮获取中文摘要

Project Summary/Abstract – Data Management and Bioinformatics Core This innovative integrated systems biology application seeks to further delineate the complex host/pathogen interactions occurring at the alveolar level that lead to unsuccessful response to therapy in serious pneumonia. The SCRIPT2 Systems Biology Center is intentionally multidisciplinary, increasing the complexity of interactions and mandating a centralized approach to facilitate accomplishing the goals of this Systems Biology Center. The overall goal of the Data Management and Bioinformatics (DMBI) Core is to develop and implement new and enhanced computational resources that support SCRIPT2, and to share those resources broadly. The DMBI Core will sit at the nexus of SCRIPT2 where it will provide the tools, methods, skills and infrastructure to collect, integrate, transform, analyze and distribute the diverse data generated by the projects. The design and implementation of the DMBI Core is based on the premise that genome-centric approaches and phenome- centric approaches are both inherently scientifically limiting. Rather, a systems biology approach that gives equal weight to all data types is more likely to produce significant findings. For the first phase of SCRIPT, the DMBI Core developed a rich data infrastructure to support the integrative analysis of clinical and multi-omic data. We will build on this foundation for SCRIPT2. The DMBI Core will advance systems biology data management in three areas. First, expanding beyond existing processes for data quality assessment, automated real-time flagging of anomalies based on statistical properties of past data, external data, and public knowledge will guide subsequent inspection toward technical rigor, biological novelty, the identification of rare patient subgroups, and provide near-time identification of emerging changes in the pathologies present by recent patients. Second, we will extend our existing SCRIPT infrastructure to allow rapid cohort identification, hypothesis generation, and analytic dataset creation. We propose extending the open-source Leaf query tool to perform both person-based and sample-based queries against both the Observational Medical Outcomes Partnership (OMOP) common data model and study-specific datasets. Tracking of specimens through the various stages of experimental and computational processing – and providing estimates of future data availability – will assist project management and preparations toward future data usage and modeling. Third, the combination of clinical data with multi-omic data with newer computational tools that combine code and data creates unique data and security challenges. Expanding on our experiences in SCRIPT to raise awareness of best practices for securely using and sharing data and models while ensuring privacy, in SCRIPT2 we will streamline processes and develop novel tools and methods to ensure a rapid and fair balance between privacy and data utility.