The documentation for the Projects building block states that a project can contain several studies. I cannot find such information in MPD. While there are the tables projects and studies, I can't find any study IDs. The two tables are very similar. I'm not sure what the purpose of the table studies in MPD is. It may be an outdated version of the table projects. We should check with the MPD staff.
In general, I don't think it makes sense to try to represent all information from MPD in EMS packets. In my opinion, it is sufficient if data contained in the package is linked to its source, e.g. by a DataSource message, which can refer to institutions, databases, publications or projects and the like. For data on individual animals, we could also include a corresponding external ID of the animal (see also issue #11).
I thought the goal of EMS packages is to bring together data from different sources in a standardized way so that different meta-analyses can be performed with greater statistical power, rather than to package all available data for individuals or cohorts. Am I right here?
The documentation for the Projects building block states that a project can contain several studies. I cannot find such information in MPD. While there are the tables
projectsandstudies, I can't find any study IDs. The two tables are very similar. I'm not sure what the purpose of the tablestudiesin MPD is. It may be an outdated version of the tableprojects. We should check with the MPD staff.In general, I don't think it makes sense to try to represent all information from MPD in EMS packets. In my opinion, it is sufficient if data contained in the package is linked to its source, e.g. by a
DataSourcemessage, which can refer to institutions, databases, publications or projects and the like. For data on individual animals, we could also include a corresponding external ID of the animal (see also issue #11).I thought the goal of EMS packages is to bring together data from different sources in a standardized way so that different meta-analyses can be performed with greater statistical power, rather than to package all available data for individuals or cohorts. Am I right here?