Automatic summarization of mouse gene information for microarray analysis by functional gene clustering and ranking of sentences in MEDLINE abstracts

With the increasing volume of published online full-text scientific articles, even the most robust Information Retrieval (IR) system returns more documents and abstracts than biomedical scientists are able to manually review. The problem is aggravated by the information-intensive nature of “high-throughput” technologies (e.g., microarray experiments) that can study expression in a given biologic context at a genome-wide scale. These advanced technologies and the increasing number of publications discussing genomic findings impair our ability to fully comprehend the meaning of the information that is embedded in the vast body of free-text biomedical literature. As such, the ability to use the literature to interpret the results of the experiments at hand is limited. Hence, tools that can survey the large quantity of literature can be helpful to the scientists interpreting and planning these large-scale genome-wide microarray experiments.