Discuss SDRF validation scope

Do we really want to validate every optional column by default??
I think we should automatically allow missing columns if they are not needed.
Sample columns are not needed if no diff. expr. is done for example.

We should do it the other way round! For specific runs, e.g. at PRIDE for some important standardized things, we should add a flag "validate_FULL_sdrf" but for general runs, it should require less things to make it easier for users.