Results of a search at SoyBase are presented in two complementary ways.
The user-submitted search term is used to find exact lexical matches to the name or description of data objects in SoyBase (QTL, genes, etc.). If an exact match is not found, and when the search term was two or more words, then a further search is done using the individual words. These text-based matches are reported in the Text column of the results table.
The goal of the ontology search is to identify data objects in SoyBase that are related to the search term but which might not have the exact search word(s) in their name or description. Data in SoyBase are tagged using a soybean-specific ontology termed a SOYnumber (SOY#). The user-submitted search term is used to search this SoyBase ontology for an exact match. If an exact match is not found, and when the search term was two or more words, the individual words are used in an additional search. Once a match is found its SOY# is used to traverse the ontology tree to identify related terms. These terms are then used to find SoyBase objects that are related to those directly identified by the search term. These ontology-based matches are reported in the Ontology column of the results table.
Occasionally unrelated SoyBase data objects will be reported in the Ontology-based results along with the correctly identified ones. This is an unavoidable consequence of the the way we have structured the underlying trait, structure and development ontologies at SoyBase. For example a search for 'protein' also returns 'oil'-related objects since both are seed components.
As an example of how these two searching capabilities can be helpful, a search for 'oil' will return those QTL that contain the word 'oil' in their name ('Seed oil', 'Seed oil to protein ratio', etc.) in the QTL line of the Text column. The Ontology column will contain these QTL along with the related QTL 'Seed palmitate', 'Seed oleic', etc.
Note: Because of the idiosyncratic ways that authors name phenotypic and developmental traits it is always advisable to look through the Ontology-based results, even when the submitted search term seems unambiguous. This can reveal additional SoyBase objects that might have been missed using exact text matching.
Funded by the USDA-ARS. Developed by the USDA-ARS SoyBase and Legume Clade Database group at the Iowa State University, Ames, IA