BioFast: Efficient and Seamless Life Science Data Management

There has been an exponential growth in the amount of data on the web that is available to the biological enterprise. This wealth of complex and diverse data presents significant opportunities and challenges for data integration and seamless access to these sources. Our research will apply prior expertise with developing data integration architectures based on wrappers and mediators to provide seamless access to heterogeneous Web accessible sources. We will develop techniques from areas such as query optimization, adaptive query evaluation, machine learning and schema mapping and integration of heterogeneous databases, to solve problems of data integration with biological data sources. Our research will address the following tasks:

• Seamless and Efficient Access: The task of query optimization becomes increasingly complex as we consider the Logical Map of Scientific classes, e.g. genes, and the Physical Map of multiple alternate sources, maultiple links between sources, and diverse query processing or domain specific computational capabilities. We extend our prior expertise of WQO (Web Query Optimization) to the biological domain to select sources and capabilities and generate efficient query execution plans. We use a simple regular expression based query language to explore the search space of alternate physical paths through the Physical Map.
• Entity Identification: A scientific entity instance may be described in multiple autonomous sources. Multi-strategy learning at the data source, schema, domain and instance level uses the concept of Identity Link to identify equivalent identifiers for the same entity instance. e.g., the gene TP53.
• Logical Link (Path) Characterization: A biologist is interested in logical links (paths) between scientific objects in the Logical Map, e.g., a sequence and a publication that studies the sequence. There may be multiple independent physical implementations of these links between different sources, in the Physical Map. We define parameterized logical links (logical paths) between data sources and define (physical) properties to describe these links. A topology of links (paths) based on their properties and concepts of Logical Link (Path) Equivalence will be developed.
• An interesting application of our research is Cancer Related Target Proteins: Identification and Complete Characterization
  

• Participants

  • Zoe Lacroix , Arizona State University, PI
  • Louiqa Raschid, University of Maryland, PI
  • Terry Gaasterland Rockefeller University (Collaborator)
  • Maria Esther Vidal, Universidad Simon Bolivar (Collaborator)
  • Felix Naumann, Humboldt University of Berlin
   

  

  Return to CLIP home page