> Do you have some examples? Biosamples represent aspects of a thing (subject/individual/organism). These aspects can be representative for the thing as a whole (e.g. genotype from germline DNA sequencing) or a part (expression profiling of a liver biopsy). This applies obviously to non-human things. While one can model environmental samples as belonging to an abstract thing (and e.g. metagenomic samples may have close relations to a specific plant, human … subject), they are primarily defined not in relation to a “biological thing” but to a combination of attributes (pacific ocean at geolocation… depth … time …). One _can_ have those in a common model, but there are certainly arguments against treating them the same. (comments here informed through lengthy discussions as GA4GH Metadata Task Team co-lead etc.; not a domain modelling specialist). Michael.
Received on Monday, 20 May 2019 16:32:44 UTC