Penta and Hexa Statements are Suitable
           for Representing Omics Knowledge
                  on the Semantic Web

                       Norio Kobayashi and Tetsuro Toyoda

       Bioinformatics And Systems Engineering division (BASE), RIKEN, JAPAN
                               toyoda@base.riken.jp??



1      Abstract

A triple or a quad in RDF is suitable to describe a relationship among a subject,
predicate and object. In a knowledge management of omics data, however, a
predicate is often used to describe a relationship between a part of subject or
object rather than whole of subject or object. For example, a statement that ‘a
feature (subject) exists within (predicate) a genomic DNA sequence (object)’ re-
quires the positional information of the feature within the genome. Therefore, an
additional URI that specifies the feature’s position within the object should be
added to the RDF statement to form a penta statement composed of five URIs
representing the statement itself, its subject, predicate, object and object’s re-
gion that specifies the position within the object. Similarly, a hexa statement
composed of six URIs representing the statement, its subject, subject’s region,
predicate, object and object’s region is conceivable as a useful representation
to associate a part of a subject with a part of an object. Here we show penta
and hexa statements are suitable for representing knowledge among huge vol-
umes of omics data. We also demonstrate an alternative way of representing a
hexa statement that is defined as an extension of conventional quad RDF, in
which we define the format of the URI of subject or object as a combination
of the subject’s or object’s URI and the internal coordinates that specify the
parts within the subject or object, respectively. The extended RDF with penta
and hexa statements are implemented in the Semantic Web Folders (SWF for
short) system which is an information platform to integrate various life science
databases and thus providing a collaboration tool among researchers to develop
new databases. SWF is available at http://database.riken.jp.




??
     To whom correspondence should be addressed.