What settings to use in GedMatch to confirm a seventh generation relative sharing DNA?

Question

I have a relative, Nancy Ann Trigg, who is 7 generations removed from me, so I expect I have one part in 2^7 (128) of her DNA in my autosomal DNA test. I want to confirm this relation and found another who is descended from Nancy (probably around 7 generations too). Comparing our DNA using gedMatch.com one-to-one comparison, I see:

Minimum threshold size to be included in total = 200 SNPs
Mismatch-bunching Limit = 100 SNPs
Minimum segment cM to be included in total = 4.0 cM

Chr Start Location  End Location    Centimorgans (cM)   SNPs
7   151,175,924 152,756,368 4.8 362
16  50,944,010  53,233,824  5.1 619
17  51,004,474  52,944,551  4.8 555
17  74,012,247  74,838,408  4.1 328
20  57,728,230  58,795,786  4.8 344
Largest segment = 5.1 cM
Total of segments > 4 cM = 23.6 cM

For my sister:

Minimum threshold size to be included in total = 200 SNPs
Mismatch-bunching Limit = 100 SNPs
Minimum segment cM to be included in total = 4.0 cM

Chr Start Location  End Location    Centimorgans (cM)   SNPs
2   7,479,337   9,272,120   6.1 445
7   29,486,454  31,878,908  4.3 701
16  50,944,010  53,233,824  5.1 620
17  51,004,474  52,944,551  4.8 557
17  70,397,218  72,817,547  4.9 626
17  73,311,163  74,418,948  4.5 350
Largest segment = 6.1 cM
Total of segments > 4 cM = 29.7 cM

For my cousin, also a descendant from Nancy:

Minimum threshold size to be included in total = 200 SNPs
Mismatch-bunching Limit = 100 SNPs
Minimum segment cM to be included in total = 4.0 cM

Chr Start Location  End Location    Centimorgans (cM)   SNPs
1   233,085,889 234,748,871 4.3 354
2   3,608,444   5,477,854   5.4 594
14  22,410,305  24,942,018  5.8 815
17  8,775,320   9,647,621   4.0 270
17  9,842,402   11,448,590  5.1 542
22  15,905,584  16,615,305  4.3 242
22  20,318,833  22,029,269  4.4 276
Largest segment = 5.8 cM
Total of segments > 4 cM = 33.2 cM

Then I did my mom, who should be unrelated:

Minimum threshold size to be included in total = 200 SNPs
Mismatch-bunching Limit = 100 SNPs
Minimum segment cM to be included in total = 4.0 cM

Chr Start Location  End Location    Centimorgans (cM)   SNPs
7   151,239,045 152,627,034 4.2 292
11  128,853,823 130,825,871 5.8 621
12  6,415,678   8,471,737   5.1 486
14  20,572,665  21,284,915  4.1 228
14  29,602,168  31,703,963  4.0 397
20  57,636,249  58,973,822  6.2 448
Largest segment = 6.2 cM
Total of segments > 4 cM = 29.3 cM

So, is a threshold of 4.0 cM and 200 SNPs a reasonable setting for a shared ancestor about 7 generations away? If not, what are better settings to use, or is there a better approach using the Tier 1 tools on GedMatch? Given that my mom matches just as much as the three who should be related, does that mean randomly picking another person of similar ancestry (e.g. European descent) going to give similar results and thus these matches are no more than random? Perhaps I need additional distant relations to Nancy to see if we can find a DNA segment that keeps repeating?

Answer 1

Autosomal DNA is not a reliable detector beyond 5 generations. At 7 generations, you're expecting all her descendants to be 1/128th part Nancy (less than 1%), and thinking you all got the same 1%. But in reality, some descendant lines may have 2% from Nancy's mother, some lines may have 1% from Nancy's father, and others got zero or near zero. You have to be really lucky, or have a lot of participants to get a strong match with current technology.