Skip to main content

Table 1 Identifying directional relationships in MEDLINE for the chemical compound caffeine on the basis of summarized relationships after analysis of 17,145 abstracts.

From: Large-scale directional relationship extraction and resolution

ObjectName

No rel.

ND Rel.

F+

F-

Fn

R+

R-

Rn

Expected

Extracted

(-)-Adrenaline

58

22

14

4

17

0

3

17

F+

F+

(-)-Dopa

5

2

1

0

1

1

0

2

F+

None

Cyclic AMP

30

9

21

11

12

4

3

13

F+

F+

diuresis

0

0

2

0

0

0

0

0

F+

F+

Drowsiness

2

0

3

1

1

1

0

0

F -

F+

Fatigue

23

8

4

6

15

1

0

11

F-

F-

Fatty acid

10

0

1

0

6

0

0

2

F+

None

gastric acid secretion

3

3

1

0

2

0

0

1

F+

None

Glycerol

15

6

2

3

8

0

0

3

F+

F -

insomnia

16

3

3

0

5

0

0

2

F+

F+

Irritability

2

2

2

1

1

0

0

1

F+

F+

Lethargy

2

0

0

1

0

0

0

0

F-

None

lipid metabolism

0

0

0

0

1

0

0

2

F+

None

Nervousness

0

1

0

1

5

0

0

1

F+

None

Norepinephrine

61

19

10

2

18

4

0

16

F+

F+

Palpitation

5

1

0

0

1

0

0

2

F+

None

PKA

3

1

1

0

3

0

0

2

F+

None

Respiratory alkalosis

0

0

0

0

1

0

0

0

F+

None

ryanodine-sensitive calcium-release channel activity

48

18

33

2

17

5

4

15

F+

F+

vasodilation

2

0

1

2

1

0

0

0

F-

F-

vasoconstriction

5

0

3

0

2

0

0

0

F+

F+

  1. Possible directional relationships are classified as F orward (caffeine affects the object) or R everse (the object affects caffeine) and as stimulatory (+), inhibitory (-) or neutral (n). No rel = no relationship found, N.D. rel = no directional relationship found, Expected = the directional relation suggested by the summary information, Extracted = the directional relation with the most support, based on the sentences processed. Bold italic font indicates errors.