High Throughput Phenotyping #604

Gabriel-Besombes · 2024-04-18T14:23:21Z

Hackathon 2024 discussion on High Throughput Phenotyping notes

Some of the points discussed :

We may need for some more generic service than "image" maybe "media" or "datafile" (the second is more specific as it doesn't include things that aren't "data" like a materials and methods pdf document for example)
We may need some kind of "provenance" metadata. This can be at two different levels : the session (computing session, image acquisition session, etc) and/or the individual data/datafile (things like the individual(s) or sensor(s) involved in acquisition).
An uncoupled system where the file descriptions are on the app's database while the actual file is stored somewhere else has a lot of advantages
In OpenSILEX/PHIS the sensors and other devices are stored as separate entities and referenced in the provenance

PROV-O : https://www.w3.org/TR/prov-o/ for provenance modelisation (might be a bit too complicated as is)
SSN : https://www.w3.org/TR/vocab-ssn/ for sensor description

This might not be the right place to store this information. Let me know if that is so.