Abstract

This paper reports on the results of two questionnaires asking biologists about the incorporation of text-extracted entity information, specifically gene and protein names, into bioscience literature search user interfaces. Among the findings are that study participants want to see gene/protein metadata in combination with organism information; that a significant proportion would like to see gene names grouped by type (synonym, homolog, etc.), and that most participants want to see information that the system is confident about immediately, and see less certain information after taking additional action. These results inform future interface designs.