Sample identifiers: samp_name vs source_mat_ID

[mslarae13]

Understanding of how to identify a biosample vs source material vs DNA sample (analyte) vs subsamples vs a globally unique ID

Some points/components that need clarification

1. PIs and researchers typically prefer to use a samp_name that is human readable, but isn't globally unique, it's only unique within that sample set. I would recommend adding an area for a unique ID

2. source_mat_ID seems like a spot for a unique ID, but MIxS is unclear on what is a sample. Is it the DNA? Or the biosample? If MIxS is meant to identify ONLY the DNA, then this makes sense. However, if the goal for MIxS is to provide biosample metadata (the soil, not just the DNA extracted from the soil), the source_mat_ID column is confusing.

• Unclear if it's meant for biomaterial links (parent-child or field soil used in lab incubation) or to provide a unique ID of the soil sample used in the DNA extraction. Definition is unclear.

[ramonawalls]

Thank you for the ticket @mslarae13. There has been quite a lot of discussion on this topic lately, and it is important to move the discussion here to a github issue.

You are correct that the intent is to have a local identifier in samp_name and a globally unique (ideally resolvable) ID in source_mat_ID. You are also quite right about the ambiguity of what ID should go into source_mat_ID. We need to clarify this, and we need a method for linking ids for the DNA extraction and the organism or environmental material from which DNA was extracted.

Some groups at NMNH have started to use source_mat_ID for the PID for the DNA extraction, which then links to the voucher system in their system. However, not everyone has the ability to set up those links the way that NMNH does.

I am part of a project called iSamples (isample.org) that is creating infrastructure to link material sample PIDs and their metadata. However, iSamples can only serve those links if they are created and stored by someone else. Therefore, I think MIxS needs to support linking among material samples in the bio domain. We have also had some internal discussions about how to create links between samples, but those are not yet summarized in a single ticket.

I will be sure to get this issues into the agenda of one of our Compliance and Interoperability Group (CIG) calls over the next few months. You are welcome to join the calls. CIG is open to anyone who wants to contribute.