Glycan pattern search

        Davide Alocci, Julien Mariethoz and Frédérique Lisacek

                        Swiss Institute of Bioinformatics



Abstract. Glycans are branched tree-like molecules composed by building
blocks linked together by chemical bonds. The molecular structure of a glycan
can be encoded into a direct acyclic graph where each node represents a build-
ing block and each edge serves as a chemical linkage between two building
blocks. In this context RDF is a possible software solution for storing structures
and SPARQL can be directly used to perform a substructure search. Glycan pat-
tern searching is an important database feature for querying structure and exper-
imental                                                                databases.

To perform a glycan pattern search, two questions need to be solved: (i) the au-
tomatic generation of a relevant SPARQL query and (ii) the import of known
glycan structures into a triple store. First we developed a software solution that
reads a structure encoded in a widely used standard in glycomics (GlycoCT),
and inserts it into a Virtuoso triple store using an ontology that we specially de-
fined for glycan structures. Then we implemented the automatic translation of a
pattern into a SPARQL query using the same ontology.

In the end the program is presented as a web interface. The user inputs the gly-
can pattern encoded in the GlycoCT format and the software retrieves all the
matching full structures in the triple store. This software is integrated and op-
erational to search patterns in the appropriate glycan-related databases (e.g.,
SugarBindDB: sugarbind.expasy.org).