User-submitted Questions (& SESAR Answers)

Question: I am currently working with a researcher that is collecting images of nanofossils using a microscope. Is there any reason that the images of individual nanofossils would not qualify as a "sample" that could be registered and given an IGSN? The samples are from the Integrated Ocean Drilling Program, and it looks like they have given the drill holes from at least some of their cruises IGSN so these samples could be subsamples of those as well.


Answer: That is a great question. I see two alternatives how to deal with images of the fossils:

  • register them with an IGSN as they basically represent the sample
  • register them with a DOI as "data" of the sample.

I think that the registration with the IGSN is more appropriate. Basically, it is the sample that gets registered (with all the sample metadata and the relation to its parent ODP core or depth interval). The IGSN metadata in SESAR includes images, so the actual image can be uploaded and can be accessed.


The difference to registration with a DOI as "data" would be that the metadata profile would include information about the image per se such as file type, file size etc., metadata that are not part of the IGSN metadata profile. I think that the correct way would be to register the sample (the nanofossil) with IGSN (sample metadata such as classification, collection, location, etc.) and register the image as "data", linking it to the IGSN as related_identifier in the DOI metadata. But that would require two registrations - too much effort. So if it is not necessary to keep information about the image itself, then I would recommend to register the image with an IGSN.



Question: For a [groundwater] sample, does each sample get one IGSN no matter if several analytics were conducted on the sample (with multiple bottles for some analytic tests)?


The user wrote: From my groundwater sampling experience, when collecting a groundwater sample, you would want to have one sample ID (IGSN)  for each GW sample no matter how many suites of analytics are performed (or separate bottles for each suite). Normally, the same GW sample ID would be on the label for each of the analytical suite bottles since it is considered one GW sample from a single source. Even if the sample bottles are sent to separate labs, you would want to know that they are all from the same sample, source, collection date and owner.  You want all the analyte results to all be associated with that single GW sample.  (If you have a different labeling practice internally, then that sample ID would also be listed in addition to the IGSN.)  Also, does the well itself get an IGSN since multiple samples may come from the same well periodically or is the well data part of the metadata for that sample?
If you or the lab has archived an extra amount or shared some of the same GW sample, and say the UofCO asks/receives an aliquot for isotope analyses or such, then their sample portion would receive a new IGSN as an off spring to the original sample with a new analytical lab/date/owner. This can then be traced back to the "original" sample.
Does this sound correct?  What is you methodology for hierarchy of IGSN? Also for physical core, does the core as a whole get an IGSN then each core segment and sample interval investigated (lab or otherwise) get the next level of IGSNs?


Answer: You are correct, right on. Your groundwater samples should each have one IGSN that is used across all labs. This will ensure that the data acquired on this sample can be unambiguously linked and integrated, especially if the data are published in multiple articles.
A relevant criteria for the use of the IGSN is the level in the 'sample hierarchy', at which data will be integrated. If analyses of 'sample splits' such as the splits of your groundwater sample can be integrated because the sample is homogeneous, they should all carry the same IGSN. If sample splits might be hetergeneous (consider a rock sample that was split before generating a homogenized rock powder), each split should have its own IGSN and be registered as a child of the original sample to make sure it can be tracked independently of the parent. Nevertheless, these data might still be published citing the IGSN of the parent sample to make it easier to link to any other data acquired on that sample.


Parent-child relationships between samples are established by recording the 'Parent IGSN' in the metadata profile of the child. If you have parent and child objects such as a core and the core's sections, you should first register the core to get the 'Parent IGSN', then register the core sections and enter the IGSN of the parent core into the metadata field 'Parent IGSN'.
It is up to you to decide, at what granularity you want to register the various objects in a sample hierarchy. For example, in case you need to know if a sample came from a specific core section or core half, you need to register the section or the section half. If the section that a sample was taken from can be deduced from the depth in core in the sample's metadata profile, then the section might not need to be registered.