Multiple Extraction Techniques for Informative Sentences

METIS combines the PRECIS annotation tool with two information extraction systems: a set of Support Vector Machine classifiers and BioIE, which is rule-based. Starting with a single query sequence, PRECIS generates structured reports from sets of related Swiss-Prot entries. The information extraction systems can then be used to extract pertinent sentences from the biomedical literature.

To use METIS, either:	Examples:
a) enter a valid Swiss-Prot identifier	NFKB2_HUMAN
b) paste a single protein sequence in FASTA format	>RBS_TOBAC MASSVLSSAAVATRSNVAQANMVAPFTGLKS AASFPVSRKQNLDITSIASNGGRVQCMQVWP PINKKKYETLSYLPDLSQEQLLSEVEYLLKNG WVPCLEFETEHGFVYRENNKSPGYYDGRYW TMWKLPMFGCTDATQVLAEVEEAKKAYPQA WIRIIGFDNVRQVQCISFIAYKPEGY
...into the box below, and provide an E-value cut off. The software will perform a BLAST search of the sequence against the Swiss-Prot database and digest related entries to produce a report.

Sequence or identifier:	BLAST E-value cut off:
	Sequences which score above this cut off are not permitted into the annotation culling process.

eMail any comments, bugs, or suggestions to: mitchell@ebi.ac.uk