Skip to main content

Table 4 Set of PPI syntax patterns

From: Protein-protein interaction extraction with feature selection by evaluating contribution levels of groups consisting of related features

No.

PPI-Pattern

Pattern 1

P1 ∗ iVerb ∗ P2

Pattern 2

P1 ∗ iVerb ∗ by ∗ P2

Pattern 3

iVerb of ∗ P1 ∗ by ∗ P2

Pattern 4

iVerb of ∗ P1 ∗ to ∗ P2

Pattern 5

iNoun of ∗ P1 ∗ [by ∣through] ∗ P2

Pattern 6

iNoun of ∗ P1 ∗ [with ∣to∣on] ∗ P2

Pattern 7

iNoun between ∗ P1 ∗ and ∗ P2

Pattern 8

complex between ∗ P1 ∗ and ∗ P2

Pattern 9

complex of ∗ P1 ∗ and ∗ P2

Pattern 10

P1 ∗ form ∗ complex with ∗ iVerb ∗ P2

Pattern 11

P1 ∗ P2 ∗ iNoun

Pattern 12

P1 depend of P2

Pattern 13

between P1 and P2

  1. We prepared syntax patterns related to PPI based on the syntax patterns proposed by Plake et al. [8]. P1 and P2 denote the protein names appearing first and later in a sentence, respectively. iNoun and iVerb denote sets of nouns and verbs related to interaction. The number of words substituted by a wildcard ‘ ∗’ in a pattern is limited to five. After the training set was divided into subsets based on the existence of significant keywords and the structure of the sentence, these syntax patterns were applied to each subset