NIF Beta Test Site

Search the NIF Advanced Search NIF Help

Welcome to the
NIF Beta Test Site NIF Resources NIF Vocabularies NIF Tools NIF Wiki   About Searching the NIF Registering Resources
to the NIF Browse the NIF Vocabulary   Add Terms to the
NIF Vocabulary Provide Feedback on
Search Interfaces

Searching the NIF Resources

Please see NIF Help page for tutorials and examples of sample searches.

    1. Levels of Access:  The NIF provides the ability to issue queries across all of these resources through a single interface. The NIF provides different levels of access to  individual resources, depending upon the type of resource and the willingness of the resource provider to register the resource and its content to the NIF (see “Registering resources to the NIF). NIF refers to these different levels of access as Level 1, Level 2 and Level 3.

    1. Level 1: Level 1 provides access to the high level description of a resource and its general content through the NIF registry. Each of these resources has been annotated with a controlled vocabulary by a human curator. This type of access is useful for looking for different types of resources, e.g., software tools and cell lines, but generally does not provide detailed information on the specific content of the resource. For example, a resource like NeuroMorpho.org that contains reconstructions of dozens of types of neurons would be annotated at the level of “neuron” but not the individual types of neurons contained within it.
    2. Level 2: Level 2 provides access to additional content of the resource, although the data structure of the resource is not explicitly mapped. For example, the different types of neurons in NeuroMorpho.org would be accessible to search, although the database would not be queried directly. In a future version of the NIF, this type of access will be provided through automated agents that “scrape” the contents of a resource and index it according to the NIF vocabularies which update as new content is added. In the current version of the NIF, level 2 access is achieved through the creation of the NIFWeb index, where content contained in web pages associated with neuroscience resources in the NIF registry and linked to these sites is indexed.; Although not comprehensive, the web index does provide access to more detailed information on resource content than Level 1. For resources that include an existing web query interface (e.g., web-searchable databases using an HTTP GET or HTTP PUT mechanism), the fields contained in this interface can be mapped into the NIFSTD ontology and/or the result sets returned can be scraped and mapped into the NIFSTD ontology.  This provides much more concept-driven query access to their underlying content than blindly scraping their content without necessitating the additional effort and/or access rights level 3 mapping requires.
    3. Level 3: Level 3 provides access to the content and data structure of the resource through the NIF data federation. For level 3, resource providers expose their database to the NIF so that it can receive a direct query. They also map the content of their database to the NIF vocabularies so that each concept in the database is identified through the unique NIF identifier. This mapping forms the basis of the concept-based query available through one of the advanced search interfaces. A resource registered at level 3 can be queried like any database.

    2. Searching a diverse set of resources and making the search results intelligible as a major challenge.  As part of this beta test, we are currently testing several search interfaces, both simple and advanced.

    1. Simple: The simple interface is a simple keyword interface, similar to many search engines. All terms that are entered are joined by an “AND”, that is, only results which contain all the terms will be returned. Enclosing the terms in quotes will search for the exact phrase.
      1. The search results are grouped according to resource type, and can be accessed by clicking through the tabs:
        1. NIFWeb: A customized web index built from neuroscience relevant sites registered to the NIF
        2. NIF registry: Resources will be returned that match the search criteria and will be categorized according to resource type, resource host and resource description. Links to a more detailed description and to the resource on the web are provided.
        3. Literature: The titles and abstracts of articles matching the search terms will be returned. Links to the articles in Pub Med or Google Scholar are provided. Access to the article is provided through these sites and may require a subscription. Links are also provided to the Textpresso for Neuroscience tool, where users can employ many additional searching and clustering strategies to explore the literature.
        4. Neuroscience databases: Relevant search results from each of the databases registered to the NIF data federation are returned. A brief description of each database and the type of data it contains is provided. A summary of results may be viewed from the results page. A link to each resource provides access to the actual record in the database.
        5. Science.gov, etc: Search results from Science.gov, a gateway to science information provided by the US government, including clinical trial and patent information.

     

    1. Advanced: The NIF utilizes many advanced features for information retrieval and integration. Chief among these is the use of a shared vocabulary for describing and querying resources. The NIF vocabularies currently consist of over 60,000 concepts derived from community ontologies and vocabularies, and enhanced through the input of neuroscience experts. More information about the NIF vocabularies can be found here.

      Through the advanced interfaces, users can make use of the NIF vocabularies to expand or refine their search and to perform so-called “concept-based queries” (see below).

      1. String-based vs. Concept-based searches:
        The NIF has taken two different approaches for searching the NIF resources: string-based and concept-based queries. The string-based search searches for text strings, similar to most existing search engines. String-based search does not rely on any special annotation of the resource and provides broad query capability. However, like most string-based search methods, it is subject to certain types of errors, e.g., it cannot distinguish between nucleus as part of a cell and nucleus as part of the brain. As explained in the next section, however, the string-based search does take advantage of the NIF vocabularies to find related terms, thereby enhancing its utility.

        The concept-based search searches not for a particular text string, but for concepts. It relies on the annotation of terms in the NIF with unique identifiers from the NIF ontology. Because it does not rely on a particular string but rather on the meaning of a term, this type of search is called “concept-based.” In the above example, nucleus as part of cell and nucleus as part of brain each map to a different unique identifier. Concept-based search is very powerful but because it relies on annotation that is largely carried out by human annotations at this time, the amount of information that is available is small relative to the string-based searches.

        Each of these methods has its own interface. The advanced interfaces let a user compose a search by using the NIF vocabularies to include related terms such as super- and sub-classes and synonyms. The advanced interface also lets users select from several query functions, e.g., “AND”, “OR” and fuzzy search.

      2. String-based interface
        Features and use of the string-based interface are described in this section. To access the string-based interface and associated tutorials, click here.
        1. Utilizing the NIF vocabularies: The NIF vocabularies are currently organized into class hierarchies, where each concept may have one parent (i.e., superclass), one or more children (subclasses) and multiple siblings (other subclasses). Each term includes a human readable definition, synonyms, abbreviations and lexical variants. When composing a search, users can take advantage of the NIF vocabularies by including parents, children and synonyms of terms, in order to enhance the likelihood of finding relevant results. To see a tutorial on using this function to find resources on neurodegenerative disease, click here. When using the NIF vocabulary operations, the NIF treats each of these compound terms as a single concept, i.e., the search is launched for “Alzheimer’s disease” rather than “Alzheimers” or “disease”.
        2. Query functions: The NIF advanced search interface allows the user to refine a search using operations like “AND” and “OR” to combine terms, e.g., Alzheimer’s disease AND Parkinson’s disease will return results that have both of these terms; Alzheimer’s disease OR Parkinson’s disease will return results that have either of these terms. In the string-based interface, these functions are represented by the “Match ALL terms “ (=AND) and “Match ANY term” (=OR).
          1. By default, the NIF will use the “OR” condition to search for parents, children and synonyms of a single term, e.g., if the “include NIF synonyms” function is selected for Alzheimer’s disease, NIF will search for “Alzheimers disease OR Alzheimer dementia OR Alzheimer Senile Dementia OR Alzheimer Type Dementia, etc.”. 
          2. If a second term is added, NIF will use either the OR or AND condition to combine the terms, depending on which is selected.  To view a tutorial on adding additional terms to a search, click here. 
          3. Fuzzy search:  In addition to using the NIF vocabularies to provide synonyms, abbreviations and lexical variants for searching the NIF sources, a fuzzy search option is available.  This option currently only issues a search on the NIF registry.  Fuzzy search essentially looks for terms that are similar to the search term, e.g., searching for “age” with fuzzy search, will also return results for “aging”.  Using this option can sometimes return spurious results, however.  For example, searching the NIF registry for “survey” with fuzzy search enabled will return many results with the term “serve”. 
        3. Concept-based interface
          To access the concept-based interface, click here.  Information on how to use the concept-based interface is available through the site.

Department of Health and Human Services National Institutes of Health