Dag Terje Filip Endresen and Bent Skovmand. Nordic Gene Bank, P.O. Box 41, Alnarp, SE-230 53, SWEDEN
Gene banks have search catalogues where potential users can search for interesting germplasm. These germplasm collections have most of their attributes in common, but the terminology used to describe them may differ substantially. Standards for data exchange and data integration have been developed and partly implemented by gene banks. Today no single point of access to search germplasm worldwide across all gene bank collections exists. The task of combining germplasm documentation datasets from different sources are challenging and methods to improve data interoperability are needed. The NGB (Nordic region), EURISCO (Europe), SINGER (CGIAR) and GRIN (USA) catalogues provide access to the passport data of major collections. But these catalogues only integrate a small set of common descriptors describing the accession IDs and origin of the accessions. Descriptive traits are often excluded and the individual online gene bank catalogues needs to be visited to find this data type. A lack of a user friendly access to the relevant documentation on the conserved accessions is a major constraint for a wider use of the germplasm holdings in gene banks. The Nordic Gene Bank has started a project on the development of methodologies and tools for trait and gene mining in the existing germplasm data on gene bank accessions. Longitude and latitude of the germplasm site origin together with measured traits and associated environmental attributes on climate, abiotic and biotic stresses will be used to build a habitat signature or digital pattern. This pattern can then be applied to germplasm passport data from other gene banks or collections of genetic resources to predict unknown trait characters. A similar approach is today in use to predict species distribution, but these methods have not yet been applied to trait mining for passport data on gene bank collections of germplasm.