Bioprospectus: Information fusion and search to support bioproduct development

Abstract

Resumen Proper use and exploitation of biodiversity resources (bioprospecting) depends upon knowledge of different organization levels (molecular, cellular and ecosystem) of biologic and genetic resources. This relies on appropriate systematic capacities to explore different information resources such as scientific literature, compound databases, medical and biological ontologies, among others. This paper presents a prototype computational system (BIOPROSPECTUS) to support bioprospecting natural products from Colombian biodiversity with biological activity. Bioprospectus is a knowledge base information retrieval (IR) system that integrates and analyzes information from multiple sources and domains, provides query suggestions based on an expert curated ontology and offer result exploration capabilities on top of a metasearch engine. The system was developed using information exploration (IE) and natural language processing (NLP) techniques over a set of raw scientific articles, integrating additional information from a collection of external knowledge bases. Users may express their information needs by combining textual and semantic keywords in a query that can be refined through domain specific knowledge structured as an ontology. For the evaluation two quantitative measures were taken and compared to a reference system. Based on the results, the system holds great promise providing a technological foundation to identify new bio-products from the colombian biodiversity targeting sustainable development and added value of our biodiversity.