Skip Navigation


Bioinformatics Advance Access originally published online on June 30, 2005
Bioinformatics 2005 21(16):3452-3453; doi:10.1093/bioinformatics/bti559
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (Print PDF) Freely available
Right arrow All Versions of this Article:
21/16/3452    most recent
bti559v1
Right arrow Comments: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when Comments are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Xuan, W.
Right arrow Articles by Meng, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Xuan, W.
Right arrow Articles by Meng, F.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions{at}oupjournals.org

GeneInfoMiner—a web server for exploring biomedical literature using batch sequence ID

Weijian Xuan , Stanley J. Watson and Fan Meng *

Department of Psychiatry and Molecular and Behavioral Neuroscience Institute, University of Michigan Ann Arbor, MI 48109, USA

*To whom correspondence should be addressed.


   Abstract
TOP
 Abstract
INTRODUCTION
 ALGORITHM AND IMPLEMENTATION
 CONCLUSION
 REFERENCES
 

Summary: GeneInfoMiner is a web-based system for searching Medline abstracts using sequence ID lists such as GenBank accession numbers derived from high-throughput experiments. It will map query results to MeSH topics to facilitate the exploration of the biological significance of the sequence ID lists. GeneInfoMiner is based on a custom gene and protein name identification engine that can map gene and protein names to important molecular biology databases.

Availability: GeneInfoMiner is freely available over the Internet at http://brainarray.mbni.med.umich.edu/GIM.asp

Contact: mengf{at}umich.edu


   INTRODUCTION
TOP
 Abstract
 INTRODUCTION
ALGORITHM AND IMPLEMENTATION
 CONCLUSION
 REFERENCES
 
The Medline database is the most utilized database for finding biological functions for gene lists derived from high-throughput studies. However, sequence IDs such as accession numbers or UniGene IDs usually cannot be used in Medline search as only 1.09% of Medline records contain sequence ID annotations and a large portion of them are related to gene cloning rather than functional studies. Furthermore, due to the redundant nature of the GenBank, one gene can usually be represented by several sequences with different sequence IDs deposited by different research groups. This redundancy issue is greatly amplified by expressed sequence tag (EST) sequencing projects for various species. Consequently, dozens, if not hundreds, of sequences may represent the same gene under many circumstances. Nonetheless, even Medline records with accession numbers do not list all possible accession numbers that can represent the same gene. As a result, Medline searches based on existing accession number annotations in Medline records usually miss a significant percentage of functional study papers.

While there are tools that aim at filtering relevant citations, e.g. MedMiner (Tanabe et al., 1999) and PubGene (Jenssen et al., 2001), they do not have the ability to explore literature by using batch sequence ID directly. In order to search Medline using sequence ID lists effectively, we developed the GeneInfoMiner based on our gene/protein name identification engine. Our implementation establishes a translation mechanism between sequence IDs and gene/protein names or abbreviations in the Medline records. Query results are mapped onto Medical Subject Heading (MeSH) terms and grouped under the hierarchical MeSH structure. As a result, users can quickly get an overview of the main functions associated with gene or probe lists derived from high-throughput experiments and easily retrieve the relevant records.


   ALGORITHM AND IMPLEMENTATION
TOP
 Abstract
 INTRODUCTION
 ALGORITHM AND IMPLEMENTATION
CONCLUSION
 REFERENCES
 
One critical part of the system is to identify gene and protein entities in biomedical text. This is a difficult task due to the lack of naming consistency. Previous attempts can be divided into two categories: systems using statistical and machine learning models and systems using hand-craft rules along with knowledge resources. GeneInfoMiner incorporates a gene and protein name identification engine developed in our group, which uses a combination of heuristics and statistical strategies (Xuan et al., 2003). In addition to the identification of gene/protein names, our program can further map extracted gene and protein names to human Entrez Gene (previously LocusLink) IDs. If a gene name stands for multiple identifiers, we use feature context (e.g. gene descriptions and abstracts cited in the Entrez Gene database) for disambiguation. We run our gene name identification program on all Medline citations and index extracted gene entities with mapped Entrez Gene IDs.

Sequence IDs submitted to the GeneInfoMiner, such as GenBank accession numbers and Affymetrix probe IDs, will first be mapped onto human Entrez Gene IDs through our ProbeMatchDB (Wang et al., 2002). The GeneInfoMiner then retrieves all papers containing these Entrez Gene IDs using the precompiled index mentioned before. Sequence IDs that our engine could not find match in Medline are listed in a separate table with the corresponding database link(s). It should be pointed out that the GeneInfoMiner is primarily for human, mouse and rat sequence ID batch queries since both our gene/protein name identification engine and the ProbeMatchDB are targeted for these three species.

Rather than simply providing a linear list of hundreds or even thousands of Medline abstracts, we utilize the MeSH (Coletti and Bleich, 2001) to organize the search results for easier literature exploration. As of this writing, over 98% Medline citations include an average of 11 MeSH terms per citation judged to be most relevant to the main topics of each citation. Our system will count the number of Medline records associated with each MeSH term and sort results according to the number of hits for each MeSH term in descending order. The GeneInfoMiner also allows the display of MeSH term relationship for individual genes by clicking the name of the gene in the result. Consequently, it is very easy for a researcher to identify the most frequent topics associated with a sequence ID list or individual genes. Because MeSH terms are assigned using full-length literature, this topic grouping function overcomes limitations in abstract-only approaches.

Medline records associated with each topic can be displayed in a browser window with multiple sorting options, such as gene symbols, the number of submitted sequence IDs related to each abstract, journal impact factor and publication date and so on. (Fig. 1) The user can also select papers of interest based on the titles and the Medline records of the selected papers can be downloaded from PubMed or directly to citation managers such as Endnote and ProCite. The corresponding full-length papers, if available, can be obtained through the links to PubMed abstracts or the URL field in citation records. Genes and MeSH terms of interest can be examined further in a clickable table with multiple database links and in hierarchical MeSH trees, respectively.



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 1 Overview of the GeneInfoMiner. It accepts batch sequence IDs as an input, maps them to Entrez Gene entries and retrieves citations. The system further groups citations using MeSH terms and provides a series of tools that facilitate researchers to find relevant citations. (a) GeneInfoMiner mapping results; (b) MeSH treeview; (c) literature selection window; (d) link to Medline citations; (e) export selected citations to citation managers and (f) gene information and links for each selected Medline citation.

 

   CONCLUSION
TOP
 Abstract
 INTRODUCTION
 ALGORITHM AND IMPLEMENTATION
 CONCLUSION
REFERENCES
 
GeneInfoMiner is designed to simplify the task of mapping sequence ID lists derived from genome-wide high-throughput studies to the Medline database. It uses our custom gene and protein name identification engine to find relationship between sequence IDs and gene/protein names in Medline records. In addition, the ability of the GeneInfoMiner to present search results according to their MeSH hit count as well as flexible sorting functions can greatly speed up the retrieval of the relevant literature related to genome-wide high-throughput results.


   Acknowledgments
 
The authors are members of the Pritzker Neuropsychiatric Disorders Research Consortium, which is supported by the Pritzker Neuropsychiatric Disorders Research Fund L.L.C. Part of this work is also supported by National Institute on Drug Abuse R21 DA13754-01 to F.M.

Conflict of Interest: none declared.

Received on January 14, 2005; revised on May 26, 2005; accepted on June 27, 2005

   REFERENCES
TOP
 Abstract
 INTRODUCTION
 ALGORITHM AND IMPLEMENTATION
 CONCLUSION
 REFERENCES
 

    Coletti, M.H. and Bleich, H.L. (2001) Medical subject headings used to search the biomedical literature. J. Am. Med. Inform. Assoc., 8, 317–323[Free Full Text].

    Jenssen, T.K., et al. (2001) A literature network of human genes for high-throughput analysis of gene expression. Nat. Genet., 28, 21–28[CrossRef][Web of Science][Medline].

    Tanabe, L., et al. (1999) MedMiner: an internet text-mining tool for biomedical information, with application to gene expression profiling. BioTechniques, 27, 1210–1217[Web of Science][Medline].

    Wang, P., et al. (2002) ProbeMatchDB—a web database for finding equivalent probes across microarray platforms and species. Bioinformatics, 18, 488–489[Abstract/Free Full Text].

    Xuan, W., et al. (2003) Identifying gene and protein names from biological texts. Proceedings of the CSB03Stanford, CA , pp. 639–643.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 BioinformaticsHome page
W. Xuan, P. Wang, S. J. Watson, and F. Meng
Medline search engine for finding genetic markers with biological significance
Bioinformatics, September 15, 2007; 23(18): 2477 - 2484.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (Print PDF) Freely available
Right arrow All Versions of this Article:
21/16/3452    most recent
bti559v1
Right arrow Comments: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when Comments are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (2)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Xuan, W.
Right arrow Articles by Meng, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Xuan, W.
Right arrow Articles by Meng, F.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?