Table 3.

Percentages of raw relationships covered using the 100 and 200 most frequent relationship types and entities. n represents the number of distinct types or entities identified in all relationships. Thus, for example, the top 200 entities modified by genes account for 85% of all raw gene-related entities mentioned in the corpus. Similarly, the top 200 relationship types account for 80% of all raw relationships in our corpus.

Percentage of covered raw relationships	relationship types (n=1921)	entities modified by
		genes (n=1210)	drugs (n=1243)	phenotypes (n=445)
by 100 most frequent relationship types or entities	68%	77%	58%	71%
by 200 most frequent relationship types or entities	80%	85%	71%	84%