I have a number of cases where the Match has an MRCA, but Clusters with a different group of Matches who clearly have a different MRCA and/or TG. Example: A Match who has an alternate MRCA which doesn’t align with a TG. I discount the MRCA because the shared DNA segment with the Match could not have come from that MRCA. Some have a paternal TG and a maternal MRCA; some are clearly from different grandparents on the same side. I now have found two examples of such a Match who Clusters with other Matches who share the MRCA but not the TG. It is not unusual for a Match to have more than one MRCA from Colonial Virginia, but usually one is closer and the closer MRCA has a much higher probability of being the one who passed down the shared DNA segment. But “higher probability” does not mean always.
My latest example is a Match with TG [01Y36] 14.0cM (on my mother’s side], but over half of the 23 Shared Matches have TG [17D25] (on my father’s side). [17D25] is a pretty well-established TG for me with an MRCA of A0170P (my CANNADAY/HILL ancestral couple at the 6C level). So I checked my Notes and found the Match has a ThruLine to CANNADAY/HILL. That explains why the Match Clusters with other Matches with MRCAs of A0170P.
Bottom line: Although my main objective is deep Chromosome Mapping, the ultimate goal is to get the genealogy right. In this case I want to also figure out the [01Y36] MRCA, so I must remove this Match from the A0170P-[17D25] Cluster. I also have to remind myself to follow the data – the data is talking to me, I need to listen…
[19Nb] Segment-ology: WTCB Issue Alt MRCA by Jim Bartlett 20220926
Thanks for this helpful post, Jim. I think I understand, but just to clarify. I have a large cluster from a 30cM run where the line of all of the known MRCAs goes back 2-5-11-23 to ancestors 46-47. But one match with an MRCA has its line as 2-4-9-19-38 and another match has a line of 2-5-11-22-44-89-178-356. These are Ancestry matches, so I don’t have segment data or most of them. I am concluding that I need to consider and research the possibility that these matches have more than one MRCA with me, but in the meantime, impute the 2-5-11-23 sequence to the matches that are not yet traced back to an MRCA and leave the anomalous matches out of this cluster. Do I understand correctly?
eallynm, A few things:
1. I think you understand this 100% – the Cluster “tends” to focus on a single Ancestor. This is complicated by several things – one of which are all the cMs from the same Ancestor. When the MRCA is a grandparent or great grandparent, it is usually the case. As we get down to 30cM, there can be a bigger spread.
2. MRCA 356 may be a legitimate genealogy MRCA, but I’d bet there was another MRCA for this Match
3. The 2-4 Match is out of whack – I’d look for gray cells (and/or Shared Matches) for this Match and, in fact, move that Match to another Cluster that it may “fit” better. I’ve done this more and more as I’ve dropped down to 20cM. Clustering is a type of “heat map” but it’s not as exacting as segment Triangulation.
4. Yes, inpute out to 23 for the other Matches
5. Look carefully at the full Cluster… Are there a few Matches that only Match one or two other Matches in the Cluster – in this case I put the DGC Excel file showing the Clusters on Monitor #2 and scroll left and right on such “week” Matches – sometimes I run into other Clusters where they are a much better fit. And in my WTCB spreadsheet, I move that row to the better Cluster.
6. Clustering is a tool – use your judgment in some cases.
Your bottom line is very well said — and something that goes far beyond this specific case, something every genealogist should have in mind. Here is a more generic version: “Bottom line: Although my main objective is XXXX, the ultimate goal is to get the genealogy right. … I also have to remind myself to follow the data – the data is talking to me, I need to listen…”
LikeLiked by 1 person
My father’s side and mother’s side came from different countries in the mid 1800s. With different languages and customs, these lines tend to stay apart in my matches.
But within those sides, sometimes a match descends from more than one of my lines on that side. So I have situations similar to what you describe, but generally just on one side. But I also have one match descended from three of my lines on my father’s side and one on my mother’s. That last eluded me for a while because the surname was spelt the same in both languages, so the language was at first no clue.
Segment triangulation helped with some of this, but it was only finding a family book for one of my father’s lines that showed the marriage into my mother’s.
I am currently resolving far more matches via segments than without. And often need all evidence from all sources.
Some matches have only made sense once I found out patterns of historical movement – some from other matches, some from histories, many from others posting about their families in forums.
Christopher, Thanks for your feedback. I have organized all of my Matches from 23andMe, FTDNA, MyHeritage and GEDmatch into 372 Triangulated Groups. This is great for organization, but those companies are relatively poor on genealogy. I’ve found a number of MRCAs, but it’s not comprehensive or across all my lines. At AncestryDNA I have multiple valid ThruLines for each of my known Ancestors back to my 5xG grandparents. But, alas, no segment data from Ancestry. I’ve found TGs for about 900 Matches at AncestryDNA (Matches who are also at one of the other sites, but most of those are below 20cM). I have just finished Walking The Clusters Back for all of my over 20cM Matches – virtually every Match is now in one of several hundred Clusters. I’m now working to link the Clusters to TGs.
My two points are: 1. I think there is enough data among all the companies to make a lot of headway on our genealgoies and Chromosome Mapping (out to 7xG grandparents in many cases); and 2. we need integrate the data across all the companies. Jim