RIC > Researcher Profiles

Your browser does not support JavaScript or JavaScript has been disabled. Please upgrade to a newer version of the browser or enable JavaScript. You might still be able to use some of the features but some of the useful features use JavaScript. Thank you!

RIC: Research Interest Comparator

Chi, Robert, Student
Affiliation: Academia Sinica
Email: cnchi@iis.sinica.edu.tw
Home Page: http://blog.robertchi.com
New This Month | New This Year | Abstract | Selected Publications | RIC Statistics Results - FULL MEDLINE:

Abstract:
BACKGROUND: Massive text mining of the biological literature holds great promise of relating disparate information and discovering new knowledge. However, disambiguation of gene symbols is a major bottleneck. RESULTS: We developed a simple thesaurus-based disambiguation algorithm that can operate with very little training data. The thesaurus comprises the information from five human genetic databases and MeSH. The extent of the homonym problem for human gene symbols is shown to be substantial (33% of the genes in our combined thesaurus had one or more ambiguous symbols), not only because one symbol can refer to multiple genes, but also because a gene symbol can have many non-gene meanings. A test set of 52,529 Medline abstracts, containing 690 ambiguous human gene symbols taken from OMIM, was automatically generated. Overall accuracy of the disambiguation algorithm was up to 92.7% on the test set. CONCLUSION: The ambiguity of human gene symbols is substantial, not only because one symbol may denote multiple genes but particularly because many symbols have other, non-gene meanings. The proposed disambiguation approach resolves most ambiguities in our test set with high accuracy, including the important gene/not a gene decisions. The algorithm is fast and scalable, enabling gene-symbol disambiguation in massive text mining applications.

Keywords extracted from the abstract: [ eliminated words list ]

Count	Word
2.773	abstracts
3.949	accuracy
6.441	algorithm
3.893	ambiguities
3.436	ambiguity
6.475	ambiguous
2.071	applications
1.396	approach
3.064	automatically
1.470	background
1.224	biological
3.847	bottleneck
1.378	combined
2.830	comprises
1.407	conclusion
1.478	containing
0.847	data
2.393	databases
2.305	decisions
3.991	denote
1.308	developed
17.966	disambiguation
3.682	discovering
3.293	disparate
2.920	enabling
1.708	extent
2.033	fast
3.873	gene
6.000	gene-symbol
1.856	generated

Count	Word
3.114	genes
1.174	genetic
1.704	great
0.994	high
3.074	holds
5.345	homonym
1.614	human
1.237	important
1.327	including
2.771	information
1.702	knowledge
1.583	literature
1.709	little
1.299	major
4.847	massive
7.008	meanings
2.730	medline
2.804	mesh
6.038	mining
0.902	more
1.061	most
2.731	multiple
11.263	non-gene
3.426	omim
1.994	only
2.907	operate
0.926	other
1.551	overall
1.752	particularly
1.756	problem

Count	Word
2.728	promise
1.743	proposed
3.098	refer
2.615	relating
3.633	resolves
0.741	results
4.309	scalable
5.049	set
1.714	simple
4.209	substantial
8.934	symbol
14.968	symbols
1.721	taken
5.381	text
9.146	thesaurus
6.000	thesaurus-based
1.680	training

RIC Statistics:

Extraction Method: Keyword Count with Lexical Variants Added

Eliminated words list: MedlinePlus List

Weighting Method: Term Frequency * Inverse Document Frequency

Database: Medline abstracts (1967 - Present)

Publication Type: All

Score Calculation Method: Cosine Similarity Method

Sort by: Score

Submission date and time: 2-3-2007, 10:16:31

Computation time: 00:00:04

Last updated: Saturday, 03-Feb-2007 10:16:35 CST