Use cases for collection site info

[arschat]

On definitions of Tier 2, we've received a couple of requests to add the collection_site field. It is requested by the Genetic Diversity Taskforce and also Heart.

Each bionetwork though has requested that in a different format.

• Genetic diversity - geography_collectionsite_latitude_longitude
  • Description: Latitude and longitude GPS coordinates of collection site.
  • Example: 1.321, 103.849
• Heart - Country collection site / Sample collection site
  • free-text assumed

Types

We discussed the available types for this field:

• latitude, longitude in degrees as proposed by GDN
  • Available formats according to google maps
  • Decimal degrees (DD): 41.40338, 2.17403
  • Degrees, minutes, and seconds (DMS): 41°24'12.2"N 2°10'26.5"E
  • Degrees and decimal minutes (DMM): 41 24.2028, 2 10.4418
• controlled vocabulary either a granular village/ city/ country level, or another location ontology we could use
• free-text with institute name or any info on location

Discussion

Most bionetworks are interested in recording this as a potential batch effect factor, however GD is interested in recording that as a human diversity factor.

With latitude, longitude, we can also compute the distance between collection sites, while with a free-text or even a controlled vocabulary option, we can only see the different options, and not calculate the distance at a later downstream stage.

However, latitude, longitude is not as straightforward as free-text since end user will have to manually find out the degrees of the institute or city/ country provided. Also, latitude longitude units permmited have to be taken care of.

[arschat]

Proceed with free-text version in the specimen level and allow inclusion of latitude longitude (if we get many of those we could implement a separate field). @arschat todo update

[arschat]

PR in HumanCellAtlas/metadata-schema#1607