Skip to main content

Table 5 The breakdown of the indirect expressions by ‘inference’ scheme

From: New challenges for text mining: mapping between text and manually curated pathways

reaction ID

reaction type

n

1

2

3

4

5

6

a

b

a

b

a

b

      

R1

Binding

56

 

56

       

R2

Binding

20

  

3

15

   

2

 

R3

phosphorylation

1

 

1

       

+ site information

5

    

4

1

    

+ modifier information

1

      

1

   

R5

degradation

1

 

1

       

+ modifier information

3

     

1

  

2

 

R7

Binding

12

 

3

2

6

 

1

   

R8

gene expression

9

3

 

4

   

2

  

R9

gene expression

10

3

2

4

   

1

  

R10

processing

12

11

1

       
  1. The indirect expressions were subdivided by ‘inference’ scheme. Biologists often build a pathway by ‘inference’ with their biological domain knowledge from indirect expressions in the text. There are nine inference schemes when regarding indirect expressions, which indicate specific reactions. The figures are the number of sentences in each ‘inference’ scheme. Class 1a: state(s) of entity(-ies) before reaction, 1b: state(s) of entity(-ies) after reaction, 2a: function(s) of entity(-ies) before reaction, 2b: function(s) of entity(-ies) after reaction, 3a: influence of state change(s) of entity(-ies), 3b: influence of functional change(s) of entity(-ies), 4: related reaction, 5: reverse reaction, 6:characteristic of reaction.