Bioscape Searching Techniques

From irefindex
Revision as of 17:41, 3 December 2009 by PaulBoddie (talk | contribs) (New page: A number of searching techniques are applied to find textual mentions of entities or concepts, particularly those where the nature of the searching is speculative, meaning that no predefin...)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

A number of searching techniques are applied to find textual mentions of entities or concepts, particularly those where the nature of the searching is speculative, meaning that no predefined lists of search terms is used, but where certain characteristic patterns are sought after in the text.

Acronym Mentions

Mentions of acronym definition phrases involve the detection of sentences containing brackets ( and ), followed by a closer inspection of such sentences, applying regular expressions which look for one of the following patterns:

  • An acronym-like term (upper-case letters, digits and hyphens) followed by a parenthesis phrase (a phrase in brackets)
  • An acronym-like term in brackets, with the preceding text then being considered as the definition or explanation of the acronym

Upon identifying a possible acronym and explanation, a test is performed to attempt to match each initial (letter or number) with a word from the explanatory text. Here, although it is tempting to only take the first letter (or digit) from each word, other approaches may be necessary involving more sophisticated tokenisation. Consider the following examples:

PubMed #10639512

"In this study, we isolated and characterized the crucial gene at the breast cancer antiestrogen resistance 1 (BCAR1) locus."

breast cancer antiestrogen resistance 1 ( BCAR1 )
initials acronym
b, c, a, r, 1 BCAR1
  • initials correspond to acronym

PubMed #10484778

"Anhidrotic ectodermal dysplasia (EDA) is a human genetic disorder of impaired ectodermal appendage development."

ectodermal dysplasia ( EDA )
initials acronym
a, e, d
(not detectable in order)
EDA
  • presumed initials only correspond to acronym if reordered

PubMed #10484776

"We identified a glyoxylate reductase/hydroxypyruvate reductase (GRHPR) cDNA clone from a human liver expressed sequence tag (EST) library."

glyoxylate reductase/hydroxypyruvate reductase ( GRHPR )
initials acronym
g, r, h, p (within a word), r
(requiring word analysis)
GRHPR
  • presumed initials only correspond if words are inspected more closely
  • words must also isolated using a more sophisticated tokeniser than one which splits words using whitespace characters

Chromosome and Maplocation Mentions

To be expanded...

PubMed #10684944

"mouse chromosome 17 and to human chromosome 16p13.3"

PubMed #10639512

"chromosome 16q23.1"

PubMed #10484772

"chromosome Xp11.2"