Skip Navigation


Bioinformatics Advance Access originally published online on December 6, 2005
Bioinformatics 2006 22(3):378-380; doi:10.1093/bioinformatics/bti821
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (Print PDF) Freely available
Right arrow All Versions of this Article:
22/3/378    most recent
bti821v1
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 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 (4)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Ding, J.
Right arrow Articles by Wurtele, E. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ding, J.
Right arrow Articles by Wurtele, E. S.
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@oxfordjournals.org

PubMed Assistant: a biologist-friendly interface for enhanced PubMed search

Jing Ding 1, LaRon M. Hughes 1, Daniel Berleant 1,*, Andy W. Fulmer 3 and Eve Syrkin Wurtele 2

1Department of Electrical and Computer Engineering Iowa State University, Ames, IA 50011, USA
2Department of Genetics, Development and Cell Biology Iowa State University, Ames, IA 50011, USA
3The Procter and Gamble Company Cincinnati, OH 45252, USA

*To whom correspondence should be addressed.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
 REFERENCES
 

Summary: MEDLINE is one of the most important bibliographical information sources for biologists and medical workers. Its PubMed interface supports Boolean queries, which are potentially expressive and exact. However, PubMed is also designed to support simplicity of use at the expense of query expressiveness and exactness. Many PubMed users have never tried explicit Boolean queries. We developed a Java program, PubMed Assistant, to make literature access easier in several ways. PubMed Assistant provides an interface that efficiently displays information about the citations and includes useful functions such as keyword highlighting, export to citation managers, clickable links to Google Scholar and others that are lacking in PubMed.

Availability: PubMed Assistant and a detailed online manual are freely available at http://metnetdb.gdcb.iastate.edu/browser under a GPL (GNU General Public License).

Contact: berleant{at}iastate.edu


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
 REFERENCES
 
MEDLINE (National Library of Medicine, 2004b, http://www.nlm.nih.gov/pubs/factsheets/medline.html) is one of the most important sources of literature citations for biologists, medical workers and others. Its collection is broad, comprehensive (~15 000 000 citations collected from 4600 journals dating back to the 1960s) and up-to-date (~10 000 new citations added per week). Its web interface, PubMed (National Library of Medicine, 2004c; http://www.ncbi.nih.gov/entrez/query.fcgi), is easy to use. A user types in one or more keywords as a query and MEDLINE returns citations accordingly. However, a citation containing the keywords is not necessarily relevant to the user's interests. To find the relevant citations, the user has to sift through the retrieved results one by one. This can be tedious. Although the ‘Limits’ (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD=Limits&DB=pubmed) function of PubMed, which restricts retrieval to a specified field (i.e. title, title/abstract, MeSH, etc.) and/or a particular time frame can be helpful, Boolean queries provide great additional flexibility and expressiveness to the query system. For example, ‘dna [mh] AND crick [au] AND 1993 [dp]’ will find citations on DNA authored by Crick in 1993. Yet the use of explicit Boolean queries is hindered by the relative difficulty of composing them. In fact, many PubMed users have never tried them.

To make Boolean queries user-friendlier and more accessible to PubMed users as well as to add other useful functionalities lacking in PubMed, we developed a Java program, PubMed Assistant. An introduction is provided in the next section. Further details can be found in its online manual (http://metnetdb.gdcb.iastate.edu/browser/manual.htm).


   PROGRAM OVERVIEW
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
 REFERENCES
 
PubMed Assistant is a stand-alone Java application that, like PubMed, serves as an interface to MEDLINE. It sends queries to MEDLINE via NCBI Entrez Utilities' ESearch service, retrieves citations via the EFetch service and finds related articles with the ELink service (National Library of Medicine, 2004a, http://eutils.ncbi.nlm.nih.gov/entrez/query/static/eutils_help.html). Unlike PubMed, however, it includes (1) a visual query editor, which eases the editing of Boolean queries; (2) a specialized citation browser, which displays citations more compactly and comprehensively and makes it easier to navigate from one citation to another; (3) an automatic and interactive query refinement tool, AutoQuery and (4) a collection of handy utility tools that provide quick and easy connections to other frequently used applications. These include exporting to citation managers and one-click Google (http://www.google.com) or Google Scholar (http://scholar.google.com) searching.

A typical use scenario includes the following steps. The user

  1. searches using a simple query, which returns hundreds or even thousands of hits;
  2. sifts through the first few hits, marking some of them as relevant;
  3. invokes AutoQuery to formulate a refined query automatically;
  4. manually modifies the refined query in the visual query editor if necessary;
  5. searches PubMed again with the refined query, resulting in a set of hits that is likely to be considerably smaller and higher in quality;
  6. sifts through the improved collection of hits, marking the relevant ones and
  7. outputs the final result to a citation manager.


   VISUAL QUERY EDITOR
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
 REFERENCES
 
The difficulties that hinder wide acceptance of Boolean queries among biologists are summarized in Table 1. Also in the table are the solutions provided by the visual query editor. One notable feature is the tree representation of a query, where Boolean operators are internal nodes and keywords are terminal (leaf) nodes (Fig. 1). Operator precedence is represented intuitively as the hierarchical structure of a tree, eliminating the tedious and error-prone process of parenthesis balancing. Interpretation of a query can be done branch by branch from the top down or from the bottom up by expanding or collapsing internal nodes. Restructuring a query is done as a drag-and-drop operation on a tree, which is more naturally suited to the syntactic structure of such queries than cut-and-paste operations in a text editor. The search field options (such as title, title/abstract, MeSH, etc.) are provided in a drop-down list so that users do not have to memorize them.


View this table:
[in this window]
[in a new window]
  		Table 1 Difficulties in editing advanced Boolean queries and the solutions provided by the visual query editor

 

Figure 1
View larger version (10K):
[in this window]
[in a new window]
  		Fig. 1 The visual query editor showing the tree representation of a query. It represents the following query: [(a NOT b) OR (c NOT d)] AND [(e NOT f) OR (g NOT h)] AND [(i NOT j) OR (k NOT l)].

 
In addition to the visual editor, PubMed Assistant also has a simple query editor built into the main user interface (Fig. 2). This editor has capabilities similar to PubMed's ‘Limits’ page. It is for occasions when advanced Boolean queries are not necessary.


Figure 2
View larger version (49K):
[in this window]
[in a new window]
  		Fig. 2 The main user interface of PubMed Assistant. Functionalities are organized in three sections. The top section (query panel) is a simple query editor, which has capabilities similar to PubMed's ‘Limits’ page. A visual query editor can be opened from here to edit advanced Boolean queries. The midsection (hits table) is where a user manages the returned citations, e.g. marks relevant hits, saves the hits to or loads them from local hard disks, exports to citation managers, links to PubMed, Google and Google Scholar, etc. The bottom section (citation panel) is a specialized browser that displays a citation's title, abstract, MeSH terms, chemical list and authors. Note that some words in the abstract are highlighted, showing matches to the query terms (blue) or user specified keywords (red).

 

   SPECIALIZED BROWSER
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
 REFERENCES
 
The following features of the specialized browser are designed to help users identify relevant citations quickly and efficiently (see the citation panel in the lower part of the main interface, Fig. 2).
  • One-click (or keystroke) navigation among citations, instead of switching back and forth between the query result page and the abstract pages as in PubMed.
  • Compact and comprehensive display of a citation including title, abstract, MeSH list (if available), chemical list (if available), authors and journal information.
  • Keyword highlighting, intended to draw user attention to the most relevant locations in a citation. The highlighter uses a fuzzy (approximate) string-matching algorithm to deal with keyword spelling variations. It also utilizes a local synonym dictionary to identify acronyms and/or other abbreviations.


   AUTOQUERY
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
UTILITY TOOLS
 CONCLUSION
 REFERENCES
 
All too often, a PubMed search with two or three keywords returns hundreds or even thousands of hits, with only a small proportion relevant to the user's interest. The AutoQuery module is designed to help users quickly and interactively find the most relevant hits.

AutoQuery takes as input a list of relevant citations and optionally the original simple query. It automatically formulates a refined query using the most commonly used words (excluding stopwords) in the list of citations, combined with the original query if included. The stringency (i.e. how many words to formulate the new query with) is a user-adjustable parameter and helps determine the degree of query refinement (e.g. the reduction in hit set size). Users can interactively adjust the stringency value based on the number of returned hits. The effectiveness of refinement in terms of the relevancy to users' interest is currently under investigation.


   UTILITY TOOLS
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
CONCLUSION
 REFERENCES
 
A collection of utility tools provides convenient connections to other popular applications. These utilities include the following.

  1. Export to citation managers. Supported formats include BibTex, EndNote (http://www.endnote.com/), Ris [Procite (http://www.procite.com/) and Reference Manager (http://www.refman.com/].
  2. Quick link to PubMed. The PMIDs in the hits table (Fig. 2) are links that will open the original citations in PubMed. In addition, a user can conveniently copy the author names and/or user-selected keywords in the title or the abstract to the query panel (Fig. 2) and perform a PubMed search.
  3. Quick link to Google/Google Scholar. The ‘Cited’ column in the citation table is reserved for the ‘Cited by’ field in Google Scholar's search result, which shows how many times an article has been cited by other articles according to Google Scholar. Google Scholar is still in beta testing phase as of this writing, so no API is yet available for other applications to extract specific information. The number also provides a clickable link to a search in Google Scholar for the exact article. A user can also perform Google/Google Scholar searches within the main interface by selecting the keywords in the title or the abstract and clicking the corresponding button.


   CONCLUSION
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
REFERENCES
 
We have developed an open source Java application, PubMed Assistant, for enhanced PubMed searches. The enhancements include easier Boolean query editing, iterative query refinement and more convenient browsing and use of search results.


   Acknowledgments
 
PubMed Assistant was funded in part by Arabidopsis 2010 DBI-0209809 and by the Procter and Gamble Company.

Conflict of Interest: none declared.


   FOOTNOTES
 
Associate Editor: John Quackenbush

Received on July 11, 2005; revised on December 2, 2005; accepted on December 5, 2005

   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 PROGRAM OVERVIEW
 VISUAL QUERY EDITOR
 SPECIALIZED BROWSER
 AUTOQUERY
 UTILITY TOOLS
 CONCLUSION
 REFERENCES
 

    National Library of Medicine. (2004a) Entrez Utilities.

    National Library of Medicine. (2004b) MEDLINE.

    National Library of Medicine. (2004c) PUBMED.


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
D. J. States, A. S. Ade, Z. C. Wright, A. V. Bookvich, and B. D. Athey
MiSearch adaptive pubMed search tool
Bioinformatics, April 1, 2009; 25(7): 974 - 976.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
M. Errami, J. D. Wren, J. M. Hicks, and H. R. Garner
eTBLAST: a web server to identify expert reviewers, appropriate journals and similar publications
Nucleic Acids Res., July 13, 2007; 35(suppl_2): W12 - W15.
[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:
22/3/378    most recent
bti821v1
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 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 (4)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Ding, J.
Right arrow Articles by Wurtele, E. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Ding, J.
Right arrow Articles by Wurtele, E. S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?