Allie LogoA Search Service for Abbreviation / Long Form

HELP

What is Allie?

Allie is a search service for abbreviations and long forms utilized in Lifesciences. It provides a solution to the issue that many abbreviations are used in the literature, and polysemous or synonymous abbreviations appear frequently, making it difficult to read and understand scientific papers that are not relevant to the reader's expertise.

Quick start

  1. nput an abbreviation, a long form, or a substring of either of them into the [KEYWORDS] box. After that, click the `Search` button. You can try 'SPF' as an example by clicking the `try "SPF"` button.

    Another way of starting a search is to embed a query in a URL. The URL http://allie.dbcls.jp/short/exact/Any/SPF.html launches the search for long form candidates to the abbreviation 'SPF'. Please refer to "URL specification" for URL specification.

    Whichever you choose, a query must be at least two character long.

  2. You will get a search result if your query matches either abbreviations or long forms ('abbreviations/long forms', hereafter). If your query matches both of them, you will get a page that requires you to choose whether the query is an abbreviation (Abbreviation Search) or a long form (Long Form Search) before displaying the search result. In case of no matches, you will get a message indicating so. Each hit result contains the following information:

    • Abbreviation(s) or long form(s)

      If your query matches an abbreviation, its corresponding long forms are shown, and the corresponding abbreviations are shown if it matches a long form. In each hit result, appearance frequency of abbreviations/long forms is shown in parentheses.
    • Research Area

      This is one of the Journal Subject Terms, which are assigned by NLM to MEDLINE journals to describe the journals' overall scope. For all of the MEDLINE entries in which an abbreviation-long form pair appears, the most frequently assigned Term is used as a research area of the pair. Since Journal Subject Terms are MeSH terms, there is a link to a page showing the information, and the frequency is shown in parentheses.
    • Co-occurring abbreviations

      They are the abbreviations that co-occur with each hit abbreviation-long form pair in PubMed/MEDLINE entries. In the result page, up to three of them are shown for each hit pair, and there is a link to a page showing all of them. The appearance frequency of each abbreviation is shown in parentheses.
    • PubMed/MEDLINE information

      This shows the publication year and the title of a paper in which a hit abbreviation-long form pair appears. In the result page, one for each pair is shown, and there is a link to a page showing all of them. Each title is an anchor text linking to its corresponding PubMed/MEDLINE entry.
  3. You have the following options to do next.

    • Change hit results by Research AREAs

      There is a list box [Research AREAs] on the left side of a result page which contains research areas related to the hit result. You can dynamically narrow down the result by research areas (select "(Any)" if you don't care any research area, this is default).
    • Change the number of hit results

      You get up to 100 hit results by default. The radio button [Display Entries] on the left side of a result page is to choose the maximum number of results to be obtained, either 100 or all.
    • Change the number of hit results shown per page

      The list box [Entries Per Page] on the left side of a result page is to choose the number of hit results shown per page from the following options: 5, 10, 20, 50, or 100.
    • Move around the result pages

      You can change the current page by the buttons [Page Control] on the left side of a result page.
    • Check information of the pair of abbreviations and long forms

      Click the anchor text `>> details` at the abbreviations/long forms area, and a new page will open (an example page is here). Conceptually identical abbreviations and long forms are grouped together, and the clustered abbreviations/long forms are labeled by the representing (most frequently used) abbreviation and long form ('cluster labels', hereafter). In that page, you can check the cluster labels and their variations, and clicking a variation shows you the RDF data. In addition, you can check the following information for each PubMed/MEDLINE entry :Each title is an anchor text linking to its corresponding PubMed entry. Clicking an abbreviation shows you the same search result as you get by entering it at the top page.
      • publication year,
      • title, and
      • co-occurring abbreviations that appear with the pair in the title or the abstract.
    • Check the MeSH term information for the research area

      Click the research area, and a new page will open (an example page is here).
    • Check a list of all the co-occurring abbreviations

      Click the `>> details` button at the co-occurring abbreviations area, and a new page will open (an example page is here). In this page, you can check the following information for each co-occurring abbreviation entry:
      • co-occurring abbreviations that appear in the title or the abstract with the pair,
      • appearance frequency of each abbreviation with the pair, and
      • appearance frequency of each abbreviation.
      Clicking the hyperlink of an abbreviation shows you the same search result as you get by entering it at the top page.
    • Check a list of all the PubMed/MEDLINE information

      Click the `>> details` button at the PubMed/MEDLINE information area, and a new page will open. (an example page is here) In this page, you can check the following information for each PubMed/MEDLINE entry:
      • publication year,
      • title, and
      • co-occurring abbreviations that appear with the pair in the title or the abstract.
      Each title has a hyperlink to its corresponding PubMed entry. Clicking the hyperlink of an abbreviation shows you the same search result as you get by entering it at the top page.
    • Change the order of the hit results

      At a result page, you can sort rows in numerically (in case of the first column) or alphabetically (the others) ascending or descending order by clicking a column title of the table.

Search options

Search method

  • Exact match: Search for any abbreviations or long forms that exactly match a given query.
  • Partial match: Search for any abbreviations or long forms that partly or exactly match a given query.

Default value is exact match.

URL specification

Searches for abbreviations or long forms

Retrieves pair information

Retrieves PubMed information

Retrieves co-occurring pairs in PubMed entries

FAQs

What do 'cluster(s)' and 'appearance frequency' that appear at the top of the search result mean?

In Allie, conceptually identical abbreviations and long forms are grouped by using a dictionary and an internally developed clustering tool. In the search result, the clustered abbreviations/long forms are displayed as a single hit result and labeled by cluster label which is shown in the 'Abbreviation/Long Form' column. 'Appearance frequency' refers to the number of clustered abbreviation(s) or long form(s) that appear in a title or an abstract.

Why are there some PubMed/MEDLINE entries that do not contain the cluster label of a hit result?

The clustering method used in Allie is based on conceptual identity, so some expressional variations are accepted within a cluster. As a result, the letter strings of a cluster label may differ from the letter strings of abbreviations/long forms that appear in the PubMed/MEDLINE entry. However, in the PubMed/MEDLINE entry there should be abbreviations/long forms that are conceptually identical to the cluster label. You can check the expressional variations in the information page of the pair of abbreviations and long forms.(an example page is here).

Is there any publication concerning Allie?

Yes, and here it is.

Y. Yamamoto, A. Yamaguchi, H. Bono and T. Takagi, "Allie: a database and a search service of abbreviations and long forms.", Database, 2011:bar03.
PubMed Entry | Full text paper available

H. Ao and T. Takagi, "ALICE: an algorithm to extract abbreviations from MEDLINE.", J Am Med Inform Assoc., 2005 Sep-Oct;12(5):576-86.
PubMed Entry | Full text paper available

How to use REST services?

herePlease refer to

What is the URL of Allie SPARQL endpoint?

The URL of Allie SPARQL endpoint is "http://data.allie.dbcls.jp/sparql".

Please refer to here for sample queries.

Where is the contact point for reporting a problem or suggestions/comments?

Here is the contact point: support AT dbcls DOT rois DOT ac DOT jp.

© Copyright DBCLS