The East African region in sub-Saharan Africa (SSA) is widely considered as one of the secondary centers of diversity for sweetpotatoes [Ipomoea batatas (L.) Lam.]. Farmers in the region typically grow landraces, but hybridizations occasionally result in new genotypes. Factors such as regional conflicts, natural disasters, disease, and land pressure continually threaten the SSA sweetpotato gene pool. Despite this threat, very little updated information is easily accessible about SSA germplasm collections. Such information is valuable for purposes of management, exploration, and conservation. Using germplasm collection data from Kenya, Tanzania, and Uganda, we demonstrate how publicly available GIS-based tools, e.g., DIVA-GIS, can be used to document current collections as well as make this information easily accessible, searchable, and portable. First, collection data from each country were compiled and known collection sites were georeferenced using available gazetteers. Following data cleaning and verification, georeferenced data were then converted into a GIS-compliant format, primarily as shapefiles. All files were then copied into storage media for exchange among stakeholders. To further demonstrate the portability of the GIS database files, available World Wide Web GIS web viewers enabled real-time access to GIS files uploaded to an experimental web site. This work demonstrates that with very little expense, access to extant SSA germplasm information for sweetpotatoes can be improved using publicly available GIS tools.
Arthur Villordon, Wambui Njuguna, Simon Gichuki, Heneriko Kulembeka, Jeremiah Simon, Bernard Yada, Phinehas Tukamuhabwa, and Don LaBonte
Arthur Villordon, Wambui Njuguna, Simon Gichuki, Philip Ndolo, Heneriko Kulembeka, Simon Jeremiah, Don LaBonte, Bernard Yada, Phinehas Tukamuhabwa, and Robert Mwanga
Web-based information delivers real-time or near-real-time data to clientele and other stakeholders. Although proprietary methods are available for interactively searching and updating databases through web interfaces, these methods generally require varying costs to maintain licensing agreements. The availability of publicly available software that require minimal or flexible licensing costs provide a cost-effective alternative to institutions that are considering access to databases via a web-accessible interface. For example, if a current web server is already configured to support hypertext preprocessor (PHP) scripts and MySQL databases, all that needs to be installed is a form script to allow the searching, inserting, and deleting of records. We describe procedures, software, and other applications that we used to develop a publicly accessible web interface to an experimental database of representative sweetpotato accessions in Kenya. The web address of this database is http://www.viazivitamu.org. This site also contains links to sweetpotato collection sites in Kenya, Tanzania, and Uganda graphically shown using a public domain GIS viewer. This demonstrates that public domain web-based tools can be configured not only to support collaborative activities among researchers in various locations, but also to provide relevant data to clients and other stakeholders.
Arthur Villordon, Wambui Njuguna, Simon Gichuki, Philip Ndolo, Heneriko Kulembeka, Simon C. Jeremiah, Don LaBonte, Bernard Yada, Phinehas Tukamuhabwa, and Robert O. M. Mwanga
Detailed information on the geographic distribution of a crop is important in planning efficient germplasm conservation strategies but is often not available, particularly for minor crops. Using germplasm collection data from Kenya, Tanzania, and Uganda, we used distribution modeling to predict the distribution of sweetpotato [Ipomoea batatas L. (Lam.)] in sub-Saharan Africa. We used a consensus modeling approach using the following algorithms: genetic algorithm for rule set prediction (GARP), maximum entropy, BIOCLIM, and DOMAIN. The predicted distribution encompasses known sweetpotato production areas as well as additional areas suited for this crop species. New geographic areas where at least three models predicted presence were in Angola, Cameroon, Central African Republic, The Congo, Democratic Republic of Congo, Gabon, Ghana, Angola, Ethiopia, Mozambique, Rwanda, and the Central African Republic. This information can be used to fill gaps in current sweetpotato germplasm collections as well as to further enhance the current presence-only based distribution model. Our approach demonstrates the usefulness of considering several models in developing distribution maps.