Hi
I have just put on github my first version my addon to import match files in GRAMPS.
It s still in early alpha stage and i m not really a good programmer.
It s compatible with the new dnasegmentmap
it only work for geneanet and myheritage.
I set the Default person to be the same as the ID and the megabase as 0, I got 0 matches. I thought the ID was the person person whose Association list is checked. When I looked at code, it appears that the default personâs association list is checked against the ID person? Not sure exactly what the rule does. Does it find everyone that has a triangulated segment match that the default person and the ID person have?
When I set the ID to a different person who I had a DNA relationship, I noticed that you only allow comma as the separator. You should also allow TAB. If the line does not have a comma, you get an index error on
debut=int(L[1])
You need to check the 1st character to see if it is a valid chromosome number (and not a comment). Only process the rest of the line if the chromosome is valid.
I ran with an ID that used comma separators and still got 0 matches. I think this is because most of my Notes use a TAB separator instead of a comma (based on import from FTDNA).
i wrtoe the rule as an example of use of m addon to import DNA data.
therefore the format of data was uite regular.
I will add your suggestion ASAP.
Let me give some explantation.
To make a triangulation we need three person.
The first one is the one that make the test and this person is the Home person.
The second one is usually the person that you donât have genealogical information. It s the person you select with ID and that match the Home Person.
and we search in the database the third person that traingulate with the two others with a segment that is greater than the number of megabase you give.
Thanks for the explanation. This makes sense. And I can see how this would be useful.
Question: shouldnât the triangulation be just on the M or P side of the chromosome? That is, if the maternal side of chromosome 1 of the default person overlapped with the paternal side of chromosome 1 for the ID person, I dont think there should be any triangulation. Or am I wrong?
I think you are wrong in that case. Just consider a simple example of two people matching. My first cousin and I match in a number of places. Her mother was my fatherâs sister, so the matching segments are on her maternal chromosomes and my paternal chromosomes. In some cases we triangulate with my brother, on his paternal segments. In another case we triangulate with a more distant male cousin on his maternal segments.
But it is indeed to have false triangulations. Consider the three people in this diagram:
Each color represents a matching segment of DNA on pairs of a given chromosome. Persons A and B have a common ancestor (green), persons B and C have a common ancestor (yellow), and persons A and C have a common ancestor (red). The fact that some are maternal or paternal is not relavant. But if you were not considering that breakdown, and were looking only at the chromosome number, you might think they all match each other, which is not the case here.
(Note that if a fourth person is added to the group, then there must be a true triangulation among some three of them.)
I think the filter rule can still be useful for identifying potential triangulations.
The next step is to write a tool to update DNA association based on genealogy.
i will put on paternal side if there s only realtion on paternel side
maternal if maternal
and unknow if there s nothing or both.
and the rule will be updated to use that information
i will do that the same way that Gary did in dnasegmentmap.
Are you adding an M/P/U flag to the segment info in the Note? I dont recommend that since the SyncAssociation would create the reciprocal relation and it may be on a different side. What if in the future you find a relationship on the other side? That is why DNAsegment calculates the side on the fly, allowing for an override in the segment info.
TAB processing in Note: what I did if the segment info contained a TAB is to remove all commas and then change all TAB to commas and then split based on commas. The source that uses TAB also includes thousands comma separators.
Hi Gary
My opinion is that the SyncAssociatin should not populate the DNA relation as it s not symmetric.
one can have paternal side and the other maternal side.
If i add the M/P/O flag i won"t sare anymore the Note.
But i will think twice about that.
Perhaps in my triangulation rule i will do the same that you.
