Skip to main content

Table 4 Individual derivational variant generation rules

From: Gene Ontology synonym generation rules lead to increased performance in biomedical concept recognition

Order

Rule

Rule defined

GO terms

Example derivations

1

Single word terms

1 {NN} {JJ}

1 GO:0043066 - negative regulation of apoptosis

1 “apoptotic down regulation”

  

2 {NN} {VB}

2 GO:0023040 - signaling via ionic flux

2 “signaled via ionic flux”

2

Double word terms

1 {NN_1 NN_2} {NN_1}, {VB_2 NN_1}, {JJ_1 NN_2}, {NN_1 JJ_2}

1 GO:0048666 - neuron development

1 “neural development”, “neurotic development”, “neuronal development”

  

2 {JJ_1 NN_2} {JJ_1}, {JJ_1 JJ_2}

2 GO:0005576 - chromosomal region

chromosomal”, “chromosome region”

3

Triple word terms

1 {NN_1 NN_2 NN_3} {NN_1 NN_3}, {NN_3 NN_1}, {VB_3}

1 GO:0052386 - cell wall thickening

1 “thickened wall”, “cellthickening”, “thickens cell wall”

4

“cell part” terms

Introduce and re-order cell part terms

GO:0035452 - extrinsic component of plastid membrane

peripheral to plastid membrane”, “extrinsic to plastid membrane”

5

“sensory perception” terms

Introduce variants of the sense - “sensory perception of {NN}”

GO:0050909 - sensory perception of taste

gustory”, “gustation

6

“transcription, X-dependent” terms

Introduce variants of “transcription”

GO:0006410 - transcription, RNA-templated

“RNA-dependent reverse transcription”, “RNA-dependent RT

7

X strand annealing activity” terms

Introduce variants of “annealing”

GO:0033592 - RNA strand annealing activity

“RNA hybridization”, “hybridize

  1. The seven patterns that we generate derivational variants are presented along with examples of each. While these are presented individually, all derivational and recursive syntactic (presented in Table 3) interact at each step. The examples provided are single GO terms, but any of the constituent terms produced through the above steps will go through all derivational rules, if possible. The bolded words in the GO Term and Synonyms generated column represent the impact of the rule. The Penn Treebank part-of-speech (POS) tags are utilized below: NN = noun, VB = verb, JJ = adjective. All varying forms were converted to the basic POS tag, e.g. NNS = plural noun and were converted to NN