Gauging triple stores with actual biological data

BMC Bioinformatics

Table 1 Characteristics of the NTNU dataset.

Graph	# triples	# classes	Max # sup	Avg # sup	# relations	# relation types
cco	2503040	89526	33	7.72	461946	30
cco_tc	3170556	89526	33	7.72	1129462	30
cco_A_thaliana	356903	12578	34	9.11	22132	30
cco_A_thaliana_tc	469484	12578	34	9.11	134713	30
cco_S_cerevisae	842344	35004	34	7.99	171825	30
cco_S_cerevisae_tc	1120545	35004	34	7.99	450026	30
cco_S_pombe	406131	14584	34	8.86	39997	30
cco_S_pombe_tc	533481	14584	34	8.86	167347	30
cco_H_sapiens	836622	29187	34	8.29	121383	30
cco_H_sapience_tc	1076760	29187	34	8.29	361521	30

A list is shown of the characteristics of the 10 graphs constituting the NTNU dataset. Reported in this table are, for each graph: the number of triples, the number of classes (the basic units in CCO), the maximum number of super classes for a class in the graph (Max #sup), the number of super classes averaged over all the classes (Avg #sup), the number of relations (predicates between two classes) and the number of distinct relation types. For technical reasons the analysis of the super class statistics was performed on random selections of 10000 classes.

ISSN: 1471-2105