Skip to content


Searching Science Literature: Gopubmed, a powerful semantic search engine

Finding papers and understanding how specific results fit into the wider knowledge world has gotten much easier with the introduction of GoPubmed.  Gopubmed is a semantic (knowledge-based) search engine for Medline (a database of medicine and life science related literature).  A semantic search engine categorizes aspects of search result.  Results can then be filtered based on the categories.  Gopubmed has a very wide array of automatic categories covering practically any subset of the search results.  One can also search for specific concepts if the categories are not available. The following pictures example is a step through of the various Gopubmed features.

1. Go to www.gopubmed.org.  You will be presented with the search interface.  The left panel has information about previous searches as well as other things.  Enter a search term here.

0000@9458_4126646144a9e87c-8ad7

2. After the search completes, hits are displayed in the document tab.  Note that there are other tabs (shown clearly in the next picture).

0001@9458_4126646144eca864-34fa

3. The GoPubmed interface is extremely well designed and can present a lot of information the various tabs.  The main search panel has several tabs-Top Author, Statistics and Documents as shown here:

0002@9458_412664614588e9a0-c853

4. Clicking on Top Author shows the single most contributing author of the selected set of papers.  This is good when you are researching a new topic or need to identify the most influential person quickly.

0003@9458_4126646145b05b06-2cd0

5. The statistics tab shows various breakdowns of the papers.  Clicking on any single number will give you the set of papers containing those categories.  The next set of pictures shows publication year, publication country, publication city and publication journal:

0004@9458_4126646145dad525-9c54

0005@9458_412664614623a67f-57e5

0006@9458_4126646146480f2c-b531

6. One can also view various graphs of the publication data:

*Graph of Publication over time:

0007@9458_412664614687c65a-59d8

*Graph of publication city

0008@9458_4126646146bcdf01-e297

*Graph of collaborating authors

0009@9458_4126646146ebe636-597b

7. Going back to the search hits, one can do the standard abstract viewing and exporting to various formats various the icons at the left:

*No abstract shown

0010@9458_4126646147240796-e91c

*Abstract

0011@9458_4126646147425ed1-372c

8. Left of the main search panel is another panel.  This left panel shows quickly various automatic categories.  This is a powerful feature.  The left panel has five tabs: my search, what, who, where and when.

0012@9458_412664614988e9a0-14ce

9. The what tab is the main useful part of the left panel it contains top terms and knowledge base.  This instantly gives a researcher various filters of the resultant hits.  The number following the category gives the number of papers with that concept in it.  This is an extremely useful way to get an overview of a topic and to enable quick filtration of results.

0013@9458_4126646149cd2de4-c5a5

10. Another useful Gopubmed feature is the ability to hover the mouse over a concept and then see some quick information.  Here, a phylogenetic tree of diatoms appears showing that is a type of Algae:

0014@9458_412664614acf1358-5b11

11. Further, clicking on Algae shows various other types of algae, each of which can be clicked for more information:

0016@9458_412664614b6b5493-edfd

12. In addition to obtaining phylogenetic trees, one can view a brief description of any of the categories as well as links to Wikipedia.  Here a description of diatoms is shown:

0015@9458_412664614b1c71c7-2439

13. Trees don’t apply only to organisms but also to concepts.  For example, one can easily see the various chlorophyll categorizations and hierarchy just by hovering the mouse over the chlorophyll category and clicking ‘tree’:

0017@9458_412664614bc048d1-c96d

14. Looking at the ‘knowledge base’ and clicking on the proteins button shows a listing of the various proteins that appear in diatom papers.  Checking the box by the protein name filters the results.  This can be very useful if one is interested in a particular protein or one wants to see an overview of protein studies in diatoms.

0018@9458_412664614c2bfe7c-ddb7

15. The ‘who’ tab has a list of authors which allows one to quickly filter by authors:

0019@9458_412664614c5e0e69-6119

16. The ‘where’ tab has two categories, ‘Earth’ and ‘Journals’:

0020@9458_412664614c9c413c-fe32

17. ‘Earth’ shows one the various countries:

0021@9458_412664614cc83fb7-7342

18. ‘Journals’ shows the papers published in various journals with a predefined High-Impact journal subcategory:

0022@9458_412664614ceca864-cef8

19. The ‘when’ tab shows the various publication dates:

0023@9458_412664614d0aff9f-1d46

20. One can select specific papers and send them to the clipboard as shown in ‘my search’ tab:

0024@9458_412664614d950c84-7f24

21. The What/Who/Where/When tabs now show the concepts present in the selected papers which allows further filtration of results.

0025@9458_412664614ddc5981-371d

Summary:

Gopubmed is a very useful tool.  The interface is intuitive, the categorization covers specific biologically related topics as well as publication statistics.  This overview covered only the most useful and novel features-there are other functionalities such as linking directly to pubmed and finding related papers.  It also looks like Gopubmed has some new features planned.

Another literature search engine that shows similar functionality to Gopubmed is the Scifinder application.  While Scifinder uses a larger database encompassing physics and chemistry journals, Scifinder is not free, not as intuitive to use and does not contain as many categories.  If life science and medical literature is what is being sought, Gopubmed is a very powerful tool for finding papers and understanding the context where those papers appears.

Reblog this post [with Zemanta]
Related Posts with Thumbnails

Posted in Explanations.

Tagged with , , , .


0 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.



Some HTML is OK

or, reply to this post via trackback.