Mapping NCBI content into Chado

Unfortunately it isn’t always clear how NCBI data should map into Chado.

This section describes what to expect when running the EUtils importer.

NCBI to Chado

NCBI to Chado mappings

NCBI Database	Chado Base Table
BioSample	biomaterial
BioProject	project
Assembly	analysis (but see Chado repo for discussion on this)
NCBI Taxon	organism

Database specific mappings

NCBI database descriptions:

• Assembly
  • Analysis type
  • Is an assembly a Chado analysis or project?
  • Undecided mappings
• BioProject
  • Notes and details
  • Indirect mappings
• BioSample
  • Undecided mappings
  • Attributes
• Pubmed
• Taxon

Linked content

The EUtils admin form has a checkbox to insert linked content. This will only insert content that is directly linked to the accession you are importing.

Consider a BioProject with many BioSamples and analyses listed. If you import that BioProject and choose to include linked records, all the directly associated BioSamples and Assemblies will also be imported.

If, however, you only wanted a subset of BioSamples in the database, you could import them individually: each BioSample would link to the BioProject, but the undesired BioSamples would not be imported into Chado. If all the BioSamples of interest were listed in an Assembly project, you could import that Assembly.

Note

Pay attention to the importer preview! The preview lets you double check the correct record will be inserted into the database. It also demonstrates which additional records will be inserted if the “Insert Linked Records” box is checked.

Problematic Links

In other cases, records may be linked indirectly via Chado. For example, project and organism cannot be directly linked, but, they are indirectly linked via biomaterials. In those cases, the linked records will be inserted. The user does not need to be notified at the preview step. The exception should be documented on that database’s page in the documentation.

This is the case for the following links:

• BioProject and NCBI Taxon (linked via biomaterial table)

• Assembly and Biomaterial (linked via the project table).

Linking some content is problematic for the current release of Chado. In cases where a link cannot be made, and the content cannot be linked indirectly, then this module should not insert the content. Instead, the formatter should notify the user that those accessions should be added directly.