peer2ref: supplement

supplement


home
about
contact us
This page contains a tutorial and supplementary information to the manuscript describing peer2ref:

How to use this server

In this tutorial, we are running this example, the abstract of a paper that reports that the NF-kappaB pathway is activated in normal luminal progenitors cell and it also is required for the development of luminal type breast cancer malignancies.

To run peer2ref you need to paste some text from your manuscript (such as abstract and introduction) on the box. Then hit the "GO!" button.

If you choose to check the box run with advanced options, you can choose the broad subject(s) of your manuscript. Appropriate reviewers for your paper will be selected among authors that have published in the journals associated with the(se) subject(s). Classifying journals by subject is not a trivial problem. Here we are using a ready-made clasification from the US National Library of Medicine. It consist off 123 broad subjects that are MeSH headings.

To find out which subjects may be appropriate in this case, we type in the box "breast cancer". An "exact match" search will look for that exact string between related subjects and titles of journals. Otherwise, matches for any of the typed words will be reported.

a strict search

This search produced the subject "Neoplasms".

a single subject

A more general search like "cancer" will produce many more subjects. Alternatively, leaving blank the search box will return all the subjects unselected:

an empty search box returns all the subjects:

By clicking on a subject a new window will show all the journals classified under that subject. Some journals belong to multiple subjects.

Select or unselect the check boxes and once you are happy click on the next button. If you don't choose any subject you will receive a warning on the next page, but you still can continue with your query.

You are about to submit your query. The results will be reported when ready in a html page. There appears the link to the web page that will display your results.

submitted query with link to results

A table with potentially useful reviewers will be displayed in the next page sorted by a score value and together with keywords from their publication profiles (see details in reference 1). Links from the authors to PubMed allow to quickly assess the suitability of the suggested candidates. The results table shows also a list of keywords, ordered by decreasing importance, that resume the publishing activity of each author.

suggested reviewers

back to top

Journals & subjects

We used the U.S. NLM classification of journals in 123 broad subjects. We only considered journals published in English.

back to top

Tips & pitfalls

  • If no good keywords are detected in the input text, the system cannot work and you will get an error. Lack of keywords is due to the input text. This may happen because the input text is a too short abstract or, in the contrary, it is a too long piece of text that is unstructured: i. e. a piece of text dealing about too many (and not enough related) things. In our experience, well written abstracts practically always produce good keywords. However if your abstract does not produce keywords, it doesn't mean necessarily that it is not well written, but it may be just too short for our system to work, or it is dealing with several different concepts. In that case, it may work that you add a little bit of text from the introduction section, or you may want trying to just paste the abstract twice.

  • If the system is able to detect keywords in the text (you didn't got an error) but still cannot find appropriate reviewers, even whith no topic of research restriction, it could mean that the research is highly original. In this case, selecting paragraphs from the introduction describing the background may help (the above remarks about long unstructured input text apply here).

    If you need further assistance you can contact us.

    back to top