New addition to BioNLP - the Protein Interaction Corpus

The Protein Interaction Corpus (PICorpus) is a annotated corpus of statements about protein interactions taken from MEDLINE abstracts. This corpus is offered here in WordFreak and Genia-style embedded XML formats. It is a refactored version of the original corpus created at the Protein Design Group at the Universidad Autonoma de Madrid.

This corpus can be used for protein entity identification, relation identification, and information extraction.

You can download the corpus from the BioNLP SourceForge download page:
http://sourceforge.net/project/showfiles.php?group_id=128424&package_id=192092

For further description of the corpus, please visit the PICorpus website
at http://bionlp.sourceforge.net/PICorpus/index.shtml.