Concept annotation in the CRAFT corpus

BMC Bioinformatics

Table 1 Counts of annotations

terminology	# total annotations	average # annotations per article	median # annotations per article	minimum # annotations per article	maximum # annotations per article
ChEBI	8,137	121	94	11	486
CL	5,760	86	58	0	435
Entrez Gene	12,277	183	155	3	543
GO BP^a	16,184	241	194	14	738
GO CC	8,354/4,707^b	125/70	97/51	9/0	499/322
GO MF	4,062	61	42	2	403
NCBITaxon^c	7,449	111	91	12	378
PRO	15,594	233	207	4	704
SO^d	22,090	330	328	72	935
all	99,907	1,491^e

^aWe are still in the process of reviewing and editing the GO BP & MF annotations for the official 1.0 version release; therefore, the statistics for these will likely change. We will update annotation statistics on the project Web site as needed.
^bWe have calculated statistics for the GO CC project both with and without the annotations of cell (GO:0005623), as these account for over half of the annotations of this project. In addition to skewing these statistics, since this is such a trivial concept that is also being annotated in the CL project, users may wish to exclude these annotations for training and evaluation of systems.
^cIn addition to the hundreds of thousands of organism entries, the NCBI Taxonomy also has a small taxonomy of types of biological taxa (e.g., phylum, genus, subgenus). For the NCBI Taxonomy pass, there are also a small number of annotations of the mentions of these taxonomic concepts in the articles; however, we have excluded these in these statistics.
^dFor the SO statistics, the independent_continuant annotations (as described in the Methodology) were excluded from the analysis.
^eThe averages of the total number of annotations per article and of unique concepts per article were calculated simply by adding up the averages for each terminological annotation pass.
Counts of annotations and of average, median, minimum, and maximum counts of annotations per article for the 67 articles constituting the initial public release of the CRAFT Corpus.

ISSN: 1471-2105