Common Names and Formats

Throughout the schema definitions, a consistent use of common attributes should be enforced. The following list should serve as guidance for schema developers.

Attribute Note
id the objects ID, used for references at the level of the databas/server instance; locally unique
name a more descriptive object label/identifier
description <ul><li>a string describing aspects of the object</li><li>not to be used for a list or nested object</li></ul>
created the time the record was created, in ISO8601 (see below)
updated the time the record was updated, in ISO8601 (see below)

Date and Time Format Specifications

Date and time formats are specified as ISO8601 compatible strings, both for time points as well as for intervals and durations. An optional required granularity may be specified as part of the respective attributes’ documentations.

Units of time

Time points

The specification of a time point is given through the concatenation of

Examples

Implementations

Durations

Durations are the most common form of time intervals. They do not refer to (e.g. start or end) time points. They are indicated with a leading “P”, followed by unit delimited quantifiers. A leading “T” is required before the start of the time components. Durations do not have to be normalized; “PT50H” is equally valid as “P2T2H”. A frequent use of durations in biomedical data resources are age values, e.g. “age at diagnosis”; but also “progression free survival”, “followup” or “time to recurrence” (these are descriptive labels, which do not necessarily represent GA4GH schema use).

Examples

Time intervals

Time intervals consist of a combination of two time designators. These can be either two time points for start and end, or one time point and a leading (time point indicates end) or trailing (time point indicates start) duration. The time elements are separated by a forward slash “/”.

While such anchored time intervals represent an option to capture different time features in a single value and to avoid disconnected references, in the context of the data schema, anchored intervals will presumably be used less frequently, with a qualitative anchor (“date of diagnosis”, “time of sampling”) representing the point of reference.

Examples