Sunday, July 12, 2020

My Last AncestryDNA AutoCluster AutoTree Analysis From Genetic Affairs - Post 2

I wrote My Last AncestryDNA AutoCluster AutoTree Analysis From Genetic Affairs - Post 1 two weeks ago, and want to share what I learned and provide some conclusions.  See the Part 1 post for my AutoCluster chart of the clusters and my AutoTree chart of probably DNA match common ancestors.

1)  I did not show one of my own AutoTree charts in Part 1 one of my AutoCluster analysis.  They provide this for each person identified by Genetic Affairs as a Common Ancestor.  As noted before, they use my Ancestry tree and the trees of my DNA matches to generate these AutoTrees.  Here is an AutoTree for the ancestor with the most known DNA Matches in the Genetic Affairs analysis:

Ann Marshman (1784-1856) is my 3rd great-grandmother, who married John Richman.  There is a similar chart for John.  I am the box in light green.  There are 11 DNA matches on this chart, all descended from John and Ann (Marshman) Richman (I have minimized the identifying information).  Three of them are my 3rd cousins, 4 are my 3rd cousins 1x removed, and 4 are 3rd cousins 2x removed.  We all have my 2nd great-grandparents, James Richman and Hannah Rich, as our Common Ancestor.

2)  Genetic Affairs provides much more information to the user, and it takes time to understand what is provided and how it can be used.  The most useful information for me is the spreadsheet file that can be manipulated and evaluated.  Here is the top of my spreadsheet file for this AncestryDNA cluster analysis from 11 March 2020 (I have minimized the identifying columns):

For me, the most important Page of the spreadsheet is the "AutoClusters 1 -- 1" page, shown above.

On the spreadsheet, the DNA Match's name is in Column A, the total cM match is in Column C, the cluster number is in Column E, the Notes I have in AncestryDNA for matches are in column F, and the size of the match's tree is in Column G.  Column H and further out to the right show the Cluster Color and if each person in the cluster matches the other persons in the cluster.  I have highlighted the lines with known Common Ancestors (from ThruLines or my family knowledge) in yellow on these charts.

Cluster 1 has 26 members in it, but only 4 of them have a tree, and only two of them have a tree larger than 19 members.  I have reviewed those trees and do not see an ancestor that matches one of my ancestors.

I know who #1 on the match list is - my 1st cousin 2x removed, with my grandparents Frederick Seaver and Bessie Richmond as our common ancestors (the yellow line on the chart above).  It's probably safe to say that this cluster is on my father's side.  And there are 25 matches for whom I can't identify a common ancestor!

As you can see from the Notes in Column F, I have added information for many of the matches, mostly based on Shared Matches provided by AncestryDNA.

The next group of matches is for Clusters 2 through 5:

There are more trees for these clusters, but again none identify a common ancestor.  I keep wondering if my tree is wrong!  Have I made a mistake, or was a close ancestor misattributed?

I have no Common Ancestors identified in clusters 2 through 5, but I do have leads from AncestryDNA Shared Matches.  I have guessed that these Clusters are also on my father's side of my tree.  There is one known Common Ancestor in Cluster 5 - 5th great-grandparents (Nathaniel horton and Sarah Pray) from the White-Oatley segment of my tree.

The screen below shows Matches 55 to 83, and most of Cluster 6.  This is another large cluster (29 members), and 10 have trees, but no Common Ancestors have been found in the Match's trees or in Ancestry ThruLines.  I think these are all Seaver-Richmond side clusters.

On the spreadsheet for lines 78 to 103, is the 21 Matches in Cluster 7.  There are ThruLines for 15 of them!

Cluster 7 clearly has Common Ancestors of my great-grandparents Thomas Richmond and Julia White!

The next screen is for Matches 104 to 129, and include Clusters 8 to 13.  I have known Common Ancestors in Cluster 10 (3rd great-grandparents John Rich and Rebecca Hill) and in Cluster 13 (also third great-grandparents John Rich and Rebecca Hill - my  great-grandfather Thomas Richmond is a grandson of the Rich/Hill couple).

3)  I could go on and on here, but there is not much point to it and I have TreeSharing to do.  I hope my readers can see the capabilities of Genetic Affairs and how it can be used to identify clusters who probably have a Common Ancestor.  

4)  AncestryDNA ThruLines and MyHeritageDNA Theory of Family Relativity are the only providers who attempt to help the user find Common Ancestors using the available trees of the DNA Matches. 

It seems to me that Genetic Affairs affirms the Common Ancestors for those Matches with shared DNA of 20 cM or more.  I have over 400 AncestryDNA ThruLines, but only 75 share 20 cM or more with me.  Some of my ThruLines in this group depend on trees other than my tree and my match's trees (using the Big Ancestry Tree) so Genetic Affairs does not find all 75 of them.

5)  Some of my observations about my AncestryDNA matches:

*  With 1,490 matches over 20 cM shared DNA, I "know" the common ancestor for 75 of them (5.0%).  These are ALL my DNA Cousins!  I am related somehow to them.

*  The lack of genealogy research, reflected in the trees of these DNA matches, is the main reason.  We can't do much about the matches with private trees (although ThruLines does find some using the Big Tree) or those with no tree.  

*  This also applies to the Ancestry Big Tree - it does not contain every person born between, say, 1750 and 1950 linked to a set of parents.  Not even every American-born person!  And I've found errors in the Ancestry Big Tree for some of my brick-wall ancestors.  

*  For many with small trees (say more than 10 members), there is an opportunity to create quick and dirty trees starting with the known ancestors in the match's tree to try to find a common ancestor.  I have done some of these, but not many, using Ancestry "Potential Ancestors," and was able to figure out about 50% of the challenges, but it takes a lot of time.  

*  Genetic Affairs is hampered with AncestryDNA matches because it does not have the Ancestry Big Tree to help.  And Ancestry Member Trees, including mine, can be wrong.  I worry about my tree more than ever now!

6)  Genetic Affairs is probably most useful in grouping DNA Matches on FamilyTreeDNA and 23andMe, GEDmatch, and potentially Living DNA, into Clusters.  There are fewer user trees on those sites, and clustering can help sort out who is related to which Common Ancestor if known family members are in the DNA Match list.

Since AncestryDNA has discontinued efforts by third-party programs like Genetic Affairs from accessing the data necessary to do the Clustering work, this may be the last AncestryDNA cluster analysis I share - unless AncestryDNA develops their own analysis, which may be in the works.  Or not...

7)  I would love to hear about your experiences with Cluster programs and how far you've gone using them and how much progress you've made with identifying Common Ancestors.


Disclosure:  I have been a paid subscriber to Genetic Affairs for several months, and really appreciate Evert-Jan Blom's program capabilities and the overall services he provides.  Thank you, Evert-Jan!

The URL for this post is:

Copyright (c) 2020, Randall J. Seaver

Please comment on this post on the website by clicking the URL above and then the "Comments" link at the bottom of each post.  Share it on Twitter, Facebook, or Pinterest using the icons below.  Or contact me by email at

1 comment:

EJ Blom said...

Just read this post, thanks for the support Randy! Pity it's not working anymore for Ancestry, now only for FTDNA. Best wishes from the Netherlands