Comments and suggestions are welcome, please send an email to yayamamo AT cb DOT k DOT u-tokyo DOT ac DOT jp.
OReFiL stands for Online Resource Finder for Lifescience. It facilitates a search for online resources that are introduced in peer-reviewed papers. You can search by MeSH terms or author names in addition to free words. OReFiL extracts all URLs from MEDLINE abstracts and PubMed-indexed BioMed Central full-papers (implementation or availability sections), and indexes them with MeSH terms and author names.
A query can be any words consisting of alphanumerical characters ("a" - "z", "A" - "Z", and "0" - "9"), "-", or "_". So, any punctuation is ignored. OReFiL supports not only a boolean based search, but also a language-model based search, which utilizes probabilistic inference. Boolean-based and language-model based searches are mutually exclusive. If you submit a query without a boolean operator, such as #band (described below), OReFiL carries out a language-model based search, and with a boolean operator OReFiL carries out a boolean-based search. For example, a query for a language-model based search is: apoptosis caspase, a query for a boolean based search is: #band(apoptosis caspase). Any word can be modified using modifiers such as .mesh (described below) to specify a context of where the word appears, such as in an author name or a title/abstract.
The following modifiers can be used to specify a context. Each modifier can be added just after a word with the delimiter '.' (period). For example, if you want to search for any online resources whose corresponding MEDLINE abstracts (including titles) contain the word pathway, you can add the modifier tiab, so, the query is pathway.tiab.
Used to search for online resources whose corresponding MEDLINE title or abstract contain word w. (Ex. alzheimer.tiab)
Used to search for online resources whose corresponding MEDLINE author list contains word w. (Ex. john.auth)
Used to search for online resources by MeSH descriptor w of the corresponding MEDLINE entry. Any MeSH descriptor (major or not) can be used. (Ex. p53.mesh)
Used to search for online resources by MeSH descriptor w of the corresponding MEDLINE entry. The MeSH descriptor used is limited to major ones or their conceptual ancestors in the MeSH hierarchy. (Ex. p53.majr)
Used to search for online resources by MeSH descriptor w of the corresponding MEDLINE. The MeSH descriptor used is limited to major ones only. (Ex. p53.noexp)
Used to search for online resources by their URLs that match w. (Ex. jp.url)
Uset to search for online resources by PubMed ID w. (Ex. 17683589.uid)
OReFiL uses Lemur Toolkit (Indri). Although the query language can be learned in detail at the original website the use of four rich operators that might be frequently used is explained here. The basic rule is that many operators begin with "#" and each operator takes arguments between brackets. A combination of operators, such as #band(metabolic pathway #1(Computational Biology).majr) is also acceptable.
This is the default operator when you just type in words in the query box without any operators. OReFiL searches for query-relevant resource entries in terms of the language model. Multiple words are acceptable and the word order may matter. An example is #combine(protein protein interaction).
This is the boolean "AND" operator. An example is #band(apoptosis caspase).
This operator takes two arguments and means "filter require", that is, OReFiL searches for resource entries that match the first argument and ranks them according to the second argument as a query. An example is #filreq(Software.noexp #1(sequence analysis)). In this example, OReFiL only returns those resources whose corresponding MEDLINE entry is associated with the MeSH term "Software" and ranks them according to the query #1(sequence analysis).
This operator is to be used to search for exact phrases. You can use this operator regardless of the search models (i.e., boolean based or language-model based). For example, if you want to search for resources that contain the exact phrase "signaling pathway", you need to compose the query #1(signaling pathway).
This means our crawler got the HTTP status code 404 when it accessed the web page of the online resource. It is highly probable that the web page removed, although it might be temporal. You can hide those resource entries by checking the "Hide page-not-found entries" box just below the query box (default setting: the box is checked).
This means our crawler could not fetch the web page of the online resource when it accessed the URL (the HTTP status code was 5xx). Several cases can be the reason; the server was too busy or down, the Internet connection was congested, or the server malfunctioned. So, the web page might be accessible when you click the link.
This means our crawler could not find the title of the online resource's web page or that of what looks like it. However, it is a sign that the online resource existed at least when our crawler accessed it.
This means our crawler could not fetch the web page when it accessed the URL for a reason not listed above. The HTTP status code and the corresponding message returned to our crawler are also displayed.
MeSH descriptors (major topics only) annotated to MEDLINE abstracts in the search result and their ancestral descriptors in the MeSH hierarchy are displayed in alphabetical order. The font size of the MeSH terms reflects each term's frequency among those MEDLINE abstracts. Those terms more frequently annotated to the MEDLINE abstracts are displayed bigger, and vice versa. They also function as a filter to narrow down the search result. By clicking one of them, the browser automatically composes and fills in an appropriate query into the query box. Users can easily add or remove a MeSH term in the query box by clicking the term in the MeSH term box; you do not need to type in a complicated query to narrow down your search. After you confirm the query box, just click the "SUBMIT" button to get a narrowed down search result.
For each entry of the search result, the following information other than the title of the resource is displayed.
Publication date of the latest PubMed record of the corresponding resource.
For each online resource, average rating, number of rating and number of comments are shown. To view the comments you need to login by using OpenID (See 'How to obtain an account').
For each online resource, MeSH terms appear at the bottom. The MeSH terms shown here are descriptors of the major topics annotated to MEDLINE abstract(s) of the online resource.
The number of PubMed entries that cite the online resource's URL in titles or abstracts, or in the main text (if available) are displayed. Exact titles are shown in another page (See below).
When you click 'Show details' the following additional information are displayed.
The links to popular web page search systems (Google and Yahoo) are for checking a kind of 'page rank' of the resource entry. By clicking the link, you can see how many web pages link to the resource. Google provide the link: operator, by which you can get a list of web pages that have a link to a web page specified after the operator. For example, you can search for web pages that have a link to PubMed by filling in the query link:www.pubmed.gov. The assumption is that the higher the number of web pages that link to a particular online resource, the more reputable it is.
The links to BioMed Central, Scirus, HighWire Press, and Google Scholar return a list of papers in which the URL of the online resource is cited. So, you can quickly search for papers that cite the online resource.
The titles of PubMed entries that cite the online resource's URL in titles or abstracts, or in the case of BMC full-papers, in the main text are displayed. The titles are hyperlinks to the corresponding PubMed entries. Also, icon(s) at the right hand side of a title is a hyperlink to the corresponding full text.
This is a link to the detailed information of an online resource entry. In addition to the information on the search result page, you can write comments or read comments posted by other users if you login by using OpenID (See 'How to obtain an account')
To be precise, the method used in OReFiL's system is known as blind (or 'pseudo') relevance feedback. It raises the ranks of relevant documents that otherwise could not be raised, whose usefulness has been empirically proven. When searching for relevant documents, the system carries out a search process twice, that is, after conducting a normal search, it searches again taking into account of the highly ranked entries of the first result by automatically generating a query for the second search.
It may be because you did not write down a URL in the title/abstract or the PubMed-indexed BMC paper. Even if that is not the case, we would be glad if you would inform us about your online resource. Send us the URL and PubMed ID pairs along with a pdf file of your publication for which you would feel comfortable being indexed in OReFiL. The email address is yayamamo AT cb DOT k DOT u-tokyo DOT ac DOT jp, thanks.
The sort order can be changed in either of the following ways.
There are three options as follows.
Checking this box will display only newly added resources on the search result list.
If you register and create an account, you can rate or make a comment on a resource.
OReFiL uses DBCLS's (Database Center for Life Science) OpenID account, which can be obtained at https://openid.dbcls.jp/account/signup.
You need to input your DBCLS OpenID in the box at the right hand side of the 'DBCLS OpenID URL:' on the top gray area. After that, just click the 'Sign In' button. You may be requested to input your password by the DBCLS OpenID server.
Just click the 'Sign Out' button on the top gray area.
First you need to sign in. After that, for each resource entry, an item 'Your Rate:' will appear. The rating is on a five point scale, represented by five blank stars. By clicking on each star, you can rate a resource.
Click the 'x' button.
First you need to sign in. After that, for each resource entry, an anchor 'Add comment ...' will appear right next to the stars. Click the anchor and a comment box will show up, where you can make a comment on the resource. Comments can be submitted by clicking the 'SUBMIT' button.
Click the anchor 'Edit comment ...' just below the comments on a resource. Edit your comment and click the 'SUBMIT' button.
Click the anchor 'Edit comment ...' just below the comments on a resource. Delete your comments in the box, and click the 'SUBMIT' button.
Open a resource's detailed information page, by clicking the 'Permalink' anchor. You can read those comments made by others at the bottom of the page.
Input a word in the box provided and MeSH terms that match the substrings of your input will be shown inside the MeSH terms box.
You can obtain a search result by Atom or RSS (0.91) feeds. The feeds can be viewed in widely used browsers, such as, Windows Internet Explorer or Mozilla FireFox. In addition, you can get the feeds of a search result by the following ways.
Newly added resources are available as via Atom/RSS feeds and can be viewed in web browsers or obtained by the following ways.
PubMed entries for a resource are available as Atom/RSS feeds and can be viewed in web browsers or obtained by the following ways.
Newly added PubMed entries for a resource are available as Atom/RSS feeds and can be viewed in web browsers or obtained by the following ways.
RSS version 0.91 cannot provide information about when an online resource was last updated. To obtain this kind of information, please use Atom.
OReFiL provides REST/SOAP Web APIs. Please see Specification of the Web APIs.
OReFiL provides REST/SOAP WebAPIs. Please see Specification of the Web APIs.
You can change the language setting (English or Japanese) by the following ways.
If you do not specify a language code, OReFiL will automatically determines the language setting based on your browser's preference (ACCEPT_LANGUAGE).
Taverna 1.7.0 or earlier versions have a problem of handling SOAP messages and cannot utilize the OReFiL SOAP API. Please use the latest version.
OReFiL has been developed and maintained by Yasunori Yamamoto, PhD at the Database Center for Life Sciences (DBCLS) (A former member of Takagi Laboratory, Department of Computational Biology, University of Tokyo).