Fig. 10.

Sample audit trail for a single path to a publication that relies on a genome of interest (see inset). In this example, a JGI-internal identifier from the source JAMO record links to another JGI-internal identifier from a second resource. This second identifier, in turn, is linked through an external NCBI identifier to the citing publication, represented here by a PubMed Central identifier. The source and target resource from which each accession is discovered is stored as a triple, providing a traceable path that can explain how each resource was discovered. This is an important improvement over manual searches, as tracking the source of each individual metadata entry would be burdensome and error-prone for a human data curator. Note that a linkage to a separate external NCBI identifier forms a branch of this audit trail but does not link to any publications. These identifiers are also stored, as additional connections may be discovered at a later time.