Semantic resources project/Use Cases/List of scientific use cases

Questions

 * 1) What SCF resources (if any) do you currently use?
 * 2) Do you use SWAN?
 * 3) Have you used any Neurocommons resources?
 * 4) What other web-resources do you use for your research?
 * 5) Are any large, publicly-available datasets relevant to your results or future plans?
 * 6) How do you find reagents for your experimental work?
 * 7) How do you find relevant publications in your research area?
 * 8) How do you communicate with other members of your lab or group?

Ideas
A list of all the (rough) ideas for scientific use-cases that have been mentioned in meetings or in notes:


 * Publications -> Datasets
 * Alternate antibodies for a protein
 * Suppliers for antibodies
 * SWAN Claims about Datasets


 * Publication filtering (by reagent)
 * Protein -> antibody -> experimental use -> publication
 * "Forum style" approaches to antibody annotation -- can we link comments to particular antibodies?
 * Naming of genes, alleles, epitopes, and using them to differentiate antibodies.

Finding Useable Antibodies
A scientist starts with a question "I want to detect the presence of a particular antigen, where can I buy an antibody to facilitate this detection?" The scientist goes to a web interface, with a editable text field, and types in the name of his or her antigen (a common protein name). The software behind the web interface finds protein terms from PRO which may match (by text matching against names, alternate names, and aliases) the given query, and presents a set of disambiguating canonical references to the user if more than one exists. Once a unique protein has been selected, the web system then follows the "specificity" link from the antibody resource back to antibody terms. It presents a list of antibodies to the user, that are annotated as specific either to the protein term, to another member of the family to which the protein belongs (if any), or to modified variants of the protein (if any). The user selects one or more antibodies from the web interface, and the system then expands those entries to present "offers to sell" each antibody, classified by supplier and with links to supplier websites and datasheets.

Disambiguating Mouse Models by Allele and Protein Information
A key component of mouse model information for disease research (the "mouse model resource") is the association between a mouse strain identifier and the alleles or genetic variation information associated with that model. Mouse model strains may be associated, or "tagged," with a disease term or indicator. However, existing mouse model catalogs make the ability to search the collection of strains by genetic variant difficult or impossible.

This use case begins with a researcher who knows the the gene or genomic locus that he or she wishes to investigate, and needs to find one or more mouse strains (and suppliers from which those strains are available) that show variation at that locus. The researcher knows the name of a gene, or an allele, that he or she wishes to investigate. This is a common name however, and is not a term from a controlled vocabulary or ontology. The researcher navigates to a web interface, and enters the name of the genetic variant in the text entry field. The PRO resource is used to resolve the variant name, and multiple options are presented to the researcher if disambiguation is needed. Finally, based on the entry and disambiguation, the PRO terms are converted into corresponding mouse model terms using the links from the mouse model resource. The mouse model information is presented to the user, organized by the disease model and phenotype information of each retrieved strain.

Antibody Literature Review
A scientist starts with the question, "where has [a particular antibody] been used before?" He or she begins by typing an antibody identifier (from our antibody resource) or an antibody catalog number (from a supplier) into a web-interface text field. If a catalog number was entered, that number is resolved into an antibody reference by the web system through the "offers to supply" predicate in our antibody resource. From the antibody resource, we identify all pubmed references in which the antibody was used, and classify those uses by research method type. The list of publications is then presented hierarchically to the user, organized by research method, with links to pubmed (or other journal source).

Aggregating Comments and Usage Information
A scientist uses an antibody in an experiment, and wants to record some important feature of the experimental use in a formal way: either the antibody didn't work as initially represented (it isn't specific to the antigen annotated in the resource), or it worked for some method but not for others, or it had to be used in a particular way in order to achieve the desired effects. The scientist goes to a web-interface with a text field, and enters the antibody identifier into the field. ''[Alternately: enters a catalog number, identifying an "offer to sell" the antibody, and uses that number to identify the antibody. It's also possible that the commenting ability should be about the offer-to-sell itself?]'' The web interface resolves the entry into an antibody resource, and presents a web form allowing free-text comment on the antibody. The comment is stored, along with the author's name and the current date, as structured annotations associated with the antibody resource.

At a later time, another scientist uses the same web form to look up an antibody and to view the free-text comments associated with it. These comments are divided either qualitatively (approval and disapproval), or classified by relevant experimental method.