Final Activity Report Summary - GLOBALBIORES (Global information systems for biomedical research)
Dr Ela Hunt's fellowship project, GLOBALBIORES, aimed to enable automated data integration from web databases and develop visualisations supporting the understanding of complex biomedical data sets. Both aims have been achieved. The work on visualisation had immediate impact in the biomedical community.
A new data visualisation solution for genetics data is called VisGenome. It supports the visualisation of data used in the search for disease genes and drug targets. VisGenome has been tested in two user studies in the context of cardiovascular disease research at the University of Glasgow and is supported by a novel data scaling algorithm called CartoonPlus. Source code and a web application are publicly available and an application note on the package was published. The package is in use at Glasgow.
Data integration work had led to the development of the following solutions.
1. Integration of data from web services for e-Commerce
2. Data integration algorithms for large data sets, PORSCHE, including a benchmark allowing for the comparison of various methods XBenchMatch
3. New approaches to data integration where each user defines individually the best ways of selecting the data they require
4. Practical solution for map mashups with genome maps, BioXMash, and extension with semantics
Beside the planned aims and objectives, additional scientific results were produced via collaborations with the University of Zurich (Prof. Burkhard Stiller and PhD student Thomas Bocek, and Prof. Klaus Dittrich and Dr Patrick Ziegler) and the University of Glasgow (Prof. Rod Page and PhD student Nadia Anwar). Those developments supported the main objectives. One collaboration resulted in the development of FastSS which performs approximate search in strings and is now being tuned to be used with biological sequences in data integration. Such techniques are required in Wikipedia, see demo at http://fastss.csg.uzh.ch. Another collaboration produced web interfaces and new query techniques for biological data sets (phylogenetic trees), see http://spira.zoology.gla.ac.uk/.
A new data visualisation solution for genetics data is called VisGenome. It supports the visualisation of data used in the search for disease genes and drug targets. VisGenome has been tested in two user studies in the context of cardiovascular disease research at the University of Glasgow and is supported by a novel data scaling algorithm called CartoonPlus. Source code and a web application are publicly available and an application note on the package was published. The package is in use at Glasgow.
Data integration work had led to the development of the following solutions.
1. Integration of data from web services for e-Commerce
2. Data integration algorithms for large data sets, PORSCHE, including a benchmark allowing for the comparison of various methods XBenchMatch
3. New approaches to data integration where each user defines individually the best ways of selecting the data they require
4. Practical solution for map mashups with genome maps, BioXMash, and extension with semantics
Beside the planned aims and objectives, additional scientific results were produced via collaborations with the University of Zurich (Prof. Burkhard Stiller and PhD student Thomas Bocek, and Prof. Klaus Dittrich and Dr Patrick Ziegler) and the University of Glasgow (Prof. Rod Page and PhD student Nadia Anwar). Those developments supported the main objectives. One collaboration resulted in the development of FastSS which performs approximate search in strings and is now being tuned to be used with biological sequences in data integration. Such techniques are required in Wikipedia, see demo at http://fastss.csg.uzh.ch. Another collaboration produced web interfaces and new query techniques for biological data sets (phylogenetic trees), see http://spira.zoology.gla.ac.uk/.