I am stumped on where to put DNA match citations like this:
Paternal relationship is confirmed by an autosomal 23andme.com test match between XXX and XXX, his father (direct ancestor). Predicted relationship from 23andMe: Father, based on sharing 3538.464 cM (47.56% DNA shared) across 22 segments. Citation generated by DNA Confirmation app ( DNAconf ), version 3.05 (updated 08.May.2025) - XXX accessed 16:16, 25 August 2025 (UTC)
(I use WikiTree as well as Gramps so I just edited its DNA citation generator’s out put a bit.)
I thought about Associations, but the manual says it is for non-familial relationships like Godfather, etc. and that is the exact opposite of what I am trying to document.
I looked at a citation on a child in a family. That works great for the parents, but doesn’t show up on the child’s information and wouldn’t work anyway for first cousins, second cousins, etc.
I looked at whether there is an add-on that would just put it in the right place. The DNASegmentMapGramplet doesn’t seem to support 23andMe and 23andMe does not provide the level of data it wants.
I suppose I could make a citation at the person level, but is that the right place.
Where is the right place to log DNA matches? I also realize that the citation would need to be recorded for both sides of the match.Capturing them in Gramps is important to me because I don’t need to check them in 23andMe and GedMatch (which is I also use). They would be captured in one place, Gramps.
(Please include your Gramps version and Operating System)
Gramps 6.0.7
Python: 3.13.13
sqlite: 3.50.4
orjson: 3.11.4
LANG: en_US.UTF-8
OS: Linux
Distribution: 6.6.99-09128-g14e87a8a9b71
I think Association is the best way to use it. DNA data is specific to the relationship between 2 people.
Depending on your 23andMe subscription, you can either get the overall autosomal match or the segment info for each match. DNASegmentMap uses the segment info if available. GEDmatch also provides the segment info.
I wonder if the Child Reference Editor (accessed by double-clicking a child in the Family Editor) might not be an appropriate place for Citing sources with proof of parentage? And for conclusions about parentage.
That worked well for my parents, but it did not appear at all in the “complete individual report.” Also what about cousin matches?
I don’t really want to upgrade to Premium on 23andMe and I’m fine with citing the percent, centimorgans and segments connected. (23andMe doesn’t provide centimorgans, but I am estimating it by using this formula: 7440 x percent shared / 100 = centimorgans shared (which was suggested by the developer of the WikiTree comparison tool.)
I agree that Associations seems like a good home if I overlook the fact it was designed for non-familial relationships.
I do not see any restriction in the GEDCOM 7 standard that Associations are limited to non-family.
For example, some genealogy programs support the ROLE of twin as an association.
Any type of DNA profile or result is a Source. To that source, you attach a Citation. This citation is then linked to an Event—in this case, the ‘Paternity’ or the specific relationship between the individuals in question. If the built-in event types don’t fit, you create a custom event and give it a descriptive name.
The actual data or analysis belongs in a Note. You can attach this note to the Citation, the Event, or the Persons involved, depending on your preference. Inside the note, you can create internal Gramps links to all the objects involved to show the proven relationships.
If you download a pdf, take a screenshot or anything else from the website, that belong to the source as a media file.
Don’t overcomplicate this.
I assume the current DNA gramplet doesn’t cover this specific type of information yet, but that is a problem to be solved there, not by breaking the core logic of Gramps.
If you want to keep your research historically accurate, do not fall into the trap of creating ‘workarounds’ just to satisfy the GEDCOM regime. DNA data are not ‘Associations’; they are Sources providing evidence for relationships between human beings. Nothing more, nothing less.
Note: This text was originally written in Norwegian and translated/edited for readability by Google AI.
I hadn’t thought about a custom event. I had thought about attaching to an existing event, but didn’t come up with a solution that work across all expected cases. However, a custom event might work for all the scenarios I am considering from parent to distant cousin where I just want to document the genetic match (who knows when one of the companies will go bust … 23andMe is already in shaky ground).
The Match “Event” does have a date or can have an arbitrary date.
Thank you.
For a custom shared event, you can also relate every person in your database using a Custom Role. For example, if it is a maternal match, you could use a role name like ‘DNA proven Mother’ (or whatever fits your specific case).
By using this approach, it is very easy to migrate your data if new features or a dedicated DNA gramplet that supports this type of data are introduced later. All your data remains structured, searchable, and easy to find back, regardless of future software changes."
And yes, it is important to keep these types of data also, so remember to take a screenshot or download as pdf if its possible to download it as a file..
OOH, and remember to date the file you get, regardless of if it is a pdf or a printscreen or a jpg.
I always write something like “as accessed {date}”
And a final tip: You can add Tags to the Event and all related objects. This makes it incredibly easy to filter them out of reports or exports if you wish to keep the DNA data private while still maintaining it in your research database.
I have started a developer discussion about a data model for Gramps to store DNA test and match data. This would allow the information about the autosomal match to be recorded using dedicated forms and fields and to specify the association between individuals who have taken tests. The data model allows this information to be searched and used in reports and future visualisation gramplets which is difficult at present. The discussion is here: Gramps DNA Data Model · gramps-project/gramps · Discussion #2292 · GitHub
Wow, looks much further along than an idea. Seems fleshed out at a high level. I will be watching to see if this moves forward. I saw some testing firms mentioned by name. Hopefully, it will support 23sndMe at some level.
This would solve my needs.
Jack D. Lail
(865) 898-3331
jacklail@gmail.com
Could I suggest here that this is the type of interesting, niche, vanity project of no use to the mainstream family historian.
Why?
I have a database of approx 20000 names covering mine and my wife’s DNA tests going back 8+ years we each have about 200 matches that can be attached to an individual in the database and of those about 20 each that were available viable GEDCOM or MyHeritage in a form suitable for use in a DNA segment match In other other words <2% and <.2%.
So interesting but of little practical use in building a Family Tree
phil
