Metadata field reference

Look-up tables for Sample metadata.

GenBank-ready fields

Each field is named <source>_<name>, for example gbif_lat_lon. A field is empty for a sample unless its required parts exist. For GEOME, each value comes from the nearest record that has it (tissue, then sample, then event). GBIF fields come from the occurrence only, never from its dataset or publisher. NCBI fields pass the BioSample value through, since BioSample attributes already use GenBank’s modifier names.

Name	GEOME (geome_)	GBIF (gbif_)	NCBI (ncbi_)
lat_lon	decimalLatitude, decimalLongitude	decimalLatitude, decimalLongitude	lat_lon (a decimal pair is converted)
collection_date	yearCollected, monthCollected, dayCollected	year, month, day; else a single-date eventDate	collection_date
geo_loc_name	country, then stateProvince and locality	country, then stateProvince and locality	geo_loc_name
specimen_voucher	genbankSpecimenVoucher; else catalogNumber, with institutionCode as a prefix	institutionCode, collectionCode, catalogNumber	specimen_voucher; else genbankSpecimenVoucher
collected_by	collectorList	recordedBy	collected_by
identified_by		identifiedBy	identified_by
tissue_type	tissueType
sex	sex	sex (lowercased)	sex (lowercased)
dev_stage	lifeStage	lifeStage (lowercased)	dev_stage (lowercased)

Example values: lat_lon 17.48260 S 149.89990 W; collection_date 2006-03-18, 2009-11, or 2009; geo_loc_name French Polynesia: Moorea; specimen_voucher USNM:FISH:407263.

• Vouchers. A voucher or catalog number that is a web address or an ARK gives an empty voucher. GBIF vouchers drop a leading repeat of the institution code (USNM 419933 becomes USNM:FISH:419933). GenBank checks the institution and collection codes against NCBI BioCollections; GBIF’s collection codes (such as Fishes) often differ from NCBI’s.
• NCBI placeholders give an empty value: missing, not collected, not applicable, not provided, restricted access, not recorded, not available, unspecified, unknown, none, na, n/a, and -, including forms like missing: control sample. A sex of unknown is therefore empty.

NCBI also offers fields that insert the [BioSample=] and [BioProject=] header tags, which GenBank uses to link the record to the BioSample and BioProject:

Name	Built from	Example
ncbi_biosample	BioSample accession	SAMN29555051
ncbi_bioproject	first BioProject of the BioSample	PRJNA720393
ncbi_biosample_sra	the SRA record’s biosample	SAMN29555051
ncbi_bioproject_sra	the SRA record’s bioproject	PRJNA720393

A BioSample can belong to several BioProjects; check that ncbi_bioproject is the one for your submission. The _sra fields are filled only for samples fetched through an SRA accession.

Raw fields

Every other field in a fetched record is available as <source>_<Level>_<field>, for example geome_Tissue_tissueType, gbif_Occurrence_occurrenceID, or ncbi_BioSample_isolate.

GBIF notes

GBIF interprets the records it publishes, so check its values before submitting:

• gbif_lat_lon is empty when GBIF flags the coordinates ZERO_COORDINATE, COORDINATE_INVALID, or COORDINATE_OUT_OF_RANGE.
• Other flags keep the value but are shown as badges in the viewer. Check values flagged PRESUMED_SWAPPED_COORDINATE, PRESUMED_NEGATED_LATITUDE or _LONGITUDE, COORDINATE_REPROJECTED, COORDINATE_ROUNDED, or COUNTRY_COORDINATE_MISMATCH.
• Coordinates of sensitive species may be deliberately generalized. Check the dataGeneralizations and coordinateUncertaintyInMeters raw fields.
• GBIF’s country is its own English name (United States of America), not always the name GenBank expects (USA). Check every source’s geo_loc_name against GenBank’s country list; marine samples use ocean or sea names from the same list.
• identifiedBy and the GBIF taxon describe the museum specimen, which may differ from your identification of the sequenced tissue.

Mapping-file columns used by Compare

The Compare tab finds your mapping-file columns by name, ignoring case. The first match wins. Use set_metadata_columns() or Mapping file columns… when yours differ.

Item	Columns tried
coordinates	lat_lon; or a latitude column (lat, latitude, decimalLatitude) plus a longitude column (lon, long, longitude, decimalLongitude)
collection_date	collection_date, date, eventDate
country	country, geo_loc_name (the part before :)
locality	locality
voucher	specimen_voucher, voucher, catalogNumber
collector	collected_by, collector, recordedBy
sex	sex
dev_stage	dev_stage, life_stage, lifeStage
taxon	Taxon (always)

NCBI’s country and locality come from the two parts of geo_loc_name, and its taxon from the BioSample organism.

Voucher search fields

When linking searches GBIF for a voucher, it reads:

Record	Fields
NCBI BioSample	specimen_voucher, genbankSpecimenVoucher, materialSampleID, voucherCatalogNumber, bio_material
GEOME	genbankSpecimenVoucher, materialSampleID, voucherCatalogNumber, otherCatalogNumbers, catalogNumber, or a bare voucherCatalogNumber with the institutionID

Fetch messages

Shown in the viewer, when hovering a faded logo, or in the warning after a fetch.

Message	Meaning
BCID not found in GEOME	No GEOME record has that ID. A typo in a well-formed BCID shows up this way.
record is private or needs a GEOME login	The record belongs to a private GEOME project, which RiboPilot cannot read.
BCID not recognized by GEOME	GEOME rejected the ID as malformed.
'x' is not a GEOME BCID (...)	The value does not look like a BCID.
GBIF occurrence not found (IDs can change when a dataset is republished)	No GBIF occurrence has that gbifID. Look the specimen up again on gbif.org.
no GBIF occurrence found for NMNH EZID ...	GBIF has no specimen with that EZID, or has not indexed it yet. Try again later, or use the gbifID.
'x' is not a GBIF occurrence ID (...)	The value is not a gbifID, gbif.org link, or NMNH EZID.
BioSample ... not found, SRA accession ... not found	NCBI has no record with that accession.
SRA accession ... has no linked BioSample	Supply the BioSample accession instead.
'x' is not a BioSample or SRA accession (...)	The value is not a recognized NCBI accession.
could not reach GEOME / GBIF / NCBI (...)	No connection. Try again later with fetch_*() or Refresh all.
GEOME / GBIF / NCBI returned HTTP ...	Any other server error, often a busy server. Try again later; for NCBI, set an API key.
N possible GBIF matches for ...; not linked	Linking found several GBIF specimens for a voucher. Paste the right gbifID.
too many GBIF matches for ...; not linked	The voucher matched more than 50 GBIF records. Paste the right gbifID.
... record links to ...; kept your ID ...	A fetched record names a different ID than yours. Check which is right.