It would help our AE use cases to have e.g. genotyping, transcription profiling, cnv profiling etc together. Phil Lord and I couldn't understand the rationale for having these separate from assays. ie why not have transcription profiling assay rather than transcription profiling.
The problem comes in when more than one assay is necessary to obtain certain information. One could argue thought that this just makes it a more complicate assay. This is important throughout.
Agreed - granularity problem. A ph reading, or an OD is a simple assay - .so are the results of complex processes where a complex instrument(s) generate data. So what's the use case - representation of all possible levels of complexity/modelling consistency /finding things in OBI to annotate with - the use cases proved this latter is hard. Some consistency also needed.
nucleic acid hybridization is defined as the hyb part and not the image analysis, feature extraction, part. We could make the sub parts of the high level processes - need to be more granular and defined classes in that case 'array based transcription profiling' or sequencing based transcription profiling perhaps? Or transcription profiling is an objective, and not the assay is what? many subparts of some assay?
these could be high level processes which assays are part of, or objectives, needs discussion.
Discussed on 2016-10-24 OBI call. The issued addressed in assay harmonization discussion which is about defining the end point of an assay and what should include in an assay. See tracker:
https://sourceforge.net/p/obi/obi-terms/810/