Title: Extract Biological Information from On-line Abstracts

ABSTRACT: Jounral abstracts are available on the web. For example, MEDLINE offers access to more than 12 million abstracts of articles published on journals of life sciences. It is really tedious and time consuming to search for the relevant journal and download them from the web. Futhermore, it is almost impossible to read thourgh all relevant abstracts. Therefore it is very likely that important articles might be overlooked. We have developed a text mining algorithm that can search the web for the relevant articles, download the abstracts and perfrom statistical analysis. This process is fully automated. I will demonstrate the effectiveness of our algorithm by using 15,000 on-line abstracts about yeast.