[Computational-biology] Molecular Biology text mining tool publicly available on line.

miguel.arregui at gmail.com miguel.arregui at gmail.com
Wed Jul 20 09:46:39 EST 2005


The tool is called EBIMed and is available from the European
Bioinformatics Institute:

http://www.ebi.ac.uk/Rebholz-s rv/ebimed

EBIMed is a software system that combines document retrieval
and information extraction from Medline in a very simple and
intuitive web interface. It provides the same basic capability
to search for Abstracts as PubMed does, only enhanced with
useful annotation on protein and gen names, molecular functions,
biological processes, cellular components, drug names and
species. All the Molecular Biology terminology found in the
text is linked to the public databases such as Swissprot and
the Gen Ontology.

Users can type precise and elaborated queries to perform their
searches. The retrieved Abstracts are processed by a series of
state of the art text mining filters which task is to identify
and annotate all the relevant biological entities present in
the text. Entity co-occurrence is detected at the sentence level
and results are displayed in tabular form, giving the user a
condensed view of all the information hidden behind his/her
query. Particular sentences as well as the original Abstracts
can be accessed easily at all times.

EBIMed provides you with a summary of all the available information
contained in the Abstracts returned by your query. Why is this
important?. There are roughly 15 million Abstracts in Medline.
The average query can return a potentially big subset of them
and then you have to partially read them to select the most
valuable according to your goals. With EBIMed you have an
alternative way to select the valuable Abstracts based on the
amount of information they contain. This approach will save you
time and effort.

Enjoy your searches.



More information about the Comp-bio mailing list