Skip to content


15 tools to locate published scientific papers and data

One of the first tasks a scientist does is to read the scientific literature about a particular topic.  In the past, this could  consist of querying a given database (such as Pubmed) using a few search terms.  Then one would look at the returned citations, select from them, go to the library stacks to get the references, read the papers, obtain some more citations and then go back to the library stacks to get a new set of papers.

Today’s research tools make it much easier to find relevant papers allowing one to obtain citation data without obtaining the paper as well as being able to locate related papers without using citation data.   The current tools enhance the use of PUBMED and other databases and make it easier to locate interesting and relevant papers.

The following list consists of 15 ‘tools’ that could be used to research a particular topic.  Some tools do unique things while others have similar functionality.

1. eTBLAST 3.0
etBLAST is a text-mining application designed to find relevant articles based on the content of an abstract.  It is designed to mitigate the disadvantages of keyword pubmed searches.  eTBLAST will extract keywords from the abstract and find other abstracts with similar keywords.  One can then view the resulting citations.  eTBLAST has an uncanny ability to find related articles that would not appear from a single keyword search.

2. MedEvi
MedEvi is a powerful search tool to extract contextual sentences from the literature based on a set of keywords.  The best way to see it is to try out the example given on the website.  The idea however is that Medevi will extract whole sentences from the keywords and show you the sentences along with the keywords underlined.  The keywords could be connected in different ways in the sentences.  It is a great tool if you are looking for specific things.

3. BioSmile Web Search 2.0

Biosmile Web Search attempts to scan a query sentence, detect the grammatical structure of the sentence and then find pubmed abstracts that are based on the query sentence.  The example sentence they had works but other sentences did not.  The site has a nice interface and it is worth a look.

4. Polysearch
Polysearch according to particular pre-set patterns.  While one may think that these patterns are limiting, they are actually quite useful because the developers of this application chose useful patterns.  For example one can search for the involvement of a particular protein in disease.  The tool will then run the appropriate query in the appropriate database returning to the user a list of citations.

5. Ontology Widget

This is a tool to find an ontology (classification) for a particular term.  This is useful, for example, when trying to understand microarray data where one would like to know, generally, where in functional space a particular gene belongs.  This tool interfaces with many different ontologies.

6. NCBO

NCBO is the National Center for Biological Ontologies.  The web page has a wide variety of software tools for using ontological data as well as links to external tools.

7. Yale Image Finder
Yale image finder is a good tool to find published, open-license images.  At one level, finding many related images for a topic gives one a tool by which one can measure their own data compared to many sources of published data.  In addition, as these images are licensed under a creative commons license, it is possible to reuse them in certain contexts.  Finally, searching for published figures could be a way to find related articles; for example if one searched for atomic force microscopy one can find published images in a wide variety of papers using atomic force microscopy.  For those used to getting the output of searches as abstracts or titles obtaining images is something new but this tool could be useful in a wide variety of situations.

8. Whatizit
Whatizit is a conceptually very interesting tool.  It will take a block of text like an abstract and then find links to keywords within the abstract.  For example if the word ‘hypertension’ it will find links to hypertension from specified databases.  There are a wide variety of databases such as swissProt, Protein Interaction database, Drug database, Disease database, Pathway Wiki among others.
One disadvantage of this tool is one has to choose the database prior to the search so one would have to be looking for particular types of information.  It would add more utility if the search tool can find all links to a particular term in all database allowing a more comprehensive search.

9. Hubmed
At first glance, hubmed seems a lot like pubmed except with a faster, cleaner interface.  However there are several notable features the set it apart from pubmed.
- To obtain a collection of related articles on a subject, Hubmed will take all the articles in your saved articles and find all the related articles of them.  It will then display all found articles according to various criteria.  This is similar in functionality to a Web of Science feature (desrcibed below).
- One has the ability to tag articles.  Tagging is another word for writing custom keywords for a particular paper.
- Hubmed can also analyze an abstract of an article for keywords and then recursively find articles based on those keywords.

10. Web of Science
Web of science is a database of citations covering all the natural and social sciences.  There are a variety of searches one can do with Web of Science.  One very useful feature is the ability to take the citations of a particular paper and then find all other papers that cite at least one citation from the original paper.  This, often very long (hundreds of citations) list of citations can then be sorted based on various parameters.  The citations can then be exported to endote and fulltext found for as many of the citations as possible.  This is a very rapid way to obtain a large number of even directly and tangentially related papers.  The disadvantage of such a search though is, because it depends on citation data, will miss relevant papers if it is not cited by one one of the papers.

11. Scifinder

Scifinder’s functionality is close to that of Web of Science.  Scifinder consists of a comprehensive citation database as well as tools to retrieve and analyze those citations.  The new web-based interface is better than the previous desktop based one but seems to be a little slower than other web based interfaces.  One nice feature of Scifinder is the categorize button.  Given a set of references, one can further refine them based on a wide variety of categories.

12. Endnote X3
Although this article is about finding citations and not about citation managers, the Find Full-text feature of Endnote deserves mention.  If one has citations in an endnote database, one can click on find full-text and then endnote will search and import pdfs into the library.  These pdfs can then be opened in an external pdf reader.  This saves a tremendous amount of time.

13. Faculty of 1000
Faculty of 1000 is a website that ranks papers based on an expert reviewer.  It is divided into a Medicine and Biology section.  It is a good tool to locate highly reviewed papers.  However one feature that includes it in this list is that one can identify a specific biology category and then obtain a list of highly reviewed papers about that subject.  This complements a review paper in a journal, because papers are added to groups in faster way.

14. Science Collaboration sites
Web 2.0 tools have been co-opted by scientists to enhance collaboration.  Many web tools such as Mendeley Desktop allow scientists to share their bibliographic information.  This is a good way to find lists of papers on a topic.

15. Pubmed

Pubmed also deserves a mention as it has improved in many ways.  Two useful features are the related articles links and the direct and indirect links to full-text articles.

Reblog this post [with Zemanta]
Related Posts with Thumbnails

Posted in Software, Web Log.

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.