Table 2

Distribution of instances by frequency and percentage (%) for each meta-knowledge value
Value GENIA ST ST(%)/GENIA(%)
frequency % frequency %
Investigation 1,914 5.3 520 5.5 1.03
Analysis 6,412 17.8 1,448 15.3 0.86
Observation 12,479 34.7 3,892 41.0 1.18
Fact 2,900 8.1 570 6.0 0.75
Method 973 2.7 40 0.4 0.16
Other 11,309 31.4 3,024 31.9 1.01
L3 33,090 91.9 8,921 94.0 1.02
L2 2,148 6.0 485 5.1 0.86
L1 749 2.1 88 0.9 0.45
Positive 33,779 93.9 8,939 94.2 1.00
Negative 2,208 6.1 555 5.8 0.95
High 1,381 3.8 466 4.9 1.28
Low 322 0.9 112 1.2 1.32
Neutral 34,284 95.3 8,916 93.9 0.99
Current 35,447 98.5 9,370 98.7 1.00
Other 540 1.5 124 1.3 0.87
Event Total 35,987 100.0 9,494 100.0 1.00

The last column shows the ratio of the percentage of events assigned the indicated value in the ST corpus to the percentage in the GENIA-MK corpus. Events that had no triggers were ignored.

Miwa et al.

Miwa et al. BMC Bioinformatics 2012 13:108   doi:10.1186/1471-2105-13-108

Open Data