Add information about 1-based or 0-based data in beacon

[blankdots]

Proposed solution

Have a info in the response that specifies if the data in the beacon is 0-based or 1-based.
While the recommendation for the API is to be 0-based ga4gh-beacon/specification#251 , that might not always be the case. Hence we will add some information to the API that a beacon deployment can specify what kind of data it has.

This is GA4GH related.

DoD (Definition of Done)

infoobject contains a key that specifies if 0-base or 1-base.

Testing

Unit test and peer review.

[blankdots]

based on @teemukataja offline conversation:

• Beacon-spec says, that coordinates are 0-based
• VCF-spec says, that coordinates are 1-based
• Different datasets (there are other dataset types, not only VCF) use either 0 or 1 as the first coordinate
• European genome browsers typically use 1-based coordinate system
• North-American genome browsers typically use 0-based coordinate system
• People use both 0- and 1- based coordinate system based on their background

[blankdots]

Solved in beacon network UI with: CSCfi/beacon-network-ui@2ffc700

[teemukataja]

Three solutions come to mind:

1. Declare the file type (because the file types have specifications, and that might convey the information to the user)

{
    "datasetAlleleResponses": [
        {
            ...,
            "info": {
                "fileType": "vcf"
            }
        }
    ]
}

2. Declare the coordinate base system, regardless of file type.

{
    "datasetAlleleResponses": [
        {
            ...,
            "info": {
                "coordinateBase": 1
            }
        }
    ]
}

3. Or combine them both

{
    "datasetAlleleResponses": [
        {
            ...,
            "info": {
                "fileType": "vcf",
                "coordinateBase": 1
            }
        }
    ]
}

We could get the fileType from the input datafiles *.vcf in beacon_init, so they are inserted into the database with the metadata.

Concerns

What if a dataset contains multiple file types? Then we could use arrays instead "fileType": ["bam", "vcf"], and "coordinateBase": [0, 1] or "coordinateBase": "mixed", but I don't know if it's typical for a dataset to contain mixed filetypes and mixed coordinate base systems... Will need to investigate.