Linking to Biozon entries

Linking directly to profile pages

Linking directly to profile pages by docID is the 'native' way of accessing Biozon entries, and has the advantages of maximal speed (since docids address objects in Biozon directly) and minimal ambiguity (since docids refer to exactly one object, and that association is guaranteed to be stable). Therefore, if a docID for a given Biozon object is known, this is the preferred method.

The format of a direct Profile URL is as follows:<docid>[?type=N]

where <docid> is the object docID . N is optional, but is present if one were to cut and paste links as appear on the Biozon site. It is an internal integer identifier representing the class of object, and is present only as a shortcut to skip the step of determining an object's type upon display. Its presence makes page loading marginally faster. If N is not known, it should be left out.

Examples of use

The following are examples of properly formed URLs to Biozon profile pages:

Linking via ProfileLink

ProfileLink is a general purpose module for linking to Biozon entries using some form of external identifier used by a database incorporated into Biozon. This is most useful when the Biozon docID is not known. Using this method suffers from two potential drawbacks:

  • Various databases use their own identifier semantics, and the sets of identifiers used by various databases is not guaranteed to be orthogonal. Thus, a single identifier is not guaranteed to match a single object or type of object. Multiple matches will be displayed as a list of entries containing links to their respective profile pages.
  • Objects are not addressed directly by random database identifiers, so doing an object lookup means initiating a search over the appropriate attributes in descriptor documents, which is slightly more overhead. In practice, this overhead generally miniscule

The syntax of a ProfileLink URL allows for restricting the search space when specifying identifiers to mitigate these potential problem. As such, the syntax is as follows:

where database is the name of a database in Biozon (e.g. swissprot, pdb), datatype is the name of desired data type (e.g. protein, interaction), <identifier> is the identifier, and date is the date specifying the time context of the object display.

Only one of database or datatype may be used in a given URL, at no times may a URL have both a database and datatype. These serve to restrict the search space for possible identifiers, and guarantee that matching objects appear in a give database or are of a particular datatype. This is useful to address the potential problems listed above, but neither are required. Both of these are human-readable names, and in parsing the URL, there is some flexibility as to their spelling, capitalization, punctuation, and abbreviation. Failing a match ProfileLink will default to a general search of all identifiers. To be on the safe side, pick the database or type name from this list

The <identifier> is free form. Various databases differ in their identifier syntax. In finding matches, Profile Link searches for exact case-sensitive matches, lowercase matches, and uppercase matches, meaning that the exact case of the identifier is not significant

The date field is flexible in how it is parsed. For example, 'May 9, 1981' parses to the same thing as '1981-05-09', as does '5/9/81'. When in doubt, use ISO 8601 dates

Examples of use

The following are examples of properly formed URLs to Biozon profile pages:

Also see