BioHaikuthon

by Semantic Haiku

I

amyloid beta
protein classified with blood
coagulation

II

dna mismatch
repair protein classified
with mismatch repair

III

arsenical pump
membrane protein classified
with plasma membrane

IV

coding sequence splice
pattern evidence used in
manual assertion

V

conductive hearing
loss subclass of conductive
hearing impairment

VI

cocaine preference
subclass of preference for
addictive substance

Semantic Haiku identifies a connected subgraph in an RDF graph using SPARQL queries for the labels of the nodes and edges in the graph. Semantic Haiku then determines whether the subgraph follows the haiku syllable pattern (by dictionary lookup of the labels). The haikus published here used the UniProt SPARQL endpoint and the Bio2RDF SPARQL endpoint (with the BioPortal dataset), but the code can be used on any endpoint.

Semantic Haiku was written in two BioHackathons in Nagasaki and Tsuruoka, and the code is available at https://github.com/leechuck/semantichaiku. To read more, see the Semantic Haiku Twitter bot!