"Screens" idea =- Decision flow chart?
-----Original Message-----
From: Alister John Marsh [mailto:ajmarsh@arrrg.org]
Sent: Wednesday, April 21, 2010 6:27 PM
To: rt-sails@comcast.net; genealogy-dna@rootsweb.com
Subject: RE: [DNA] GENEALOGY-DNA Intrepreting Y-DNA Matches


I guess with genetic genealogy, it could be looked at as putting people through a hierarchical series of screens. There are no doubt more possible screens than the following, but this is an example...

  1. 1) Are two people of the same haplogroup. If not, they are unrelated in recent millennia.
  2. 2) If the two people have the same haplogroup, do they have a large or small genetic distance.
    1. Large G.D.
    2. a) If the genetic distance is large, has a RecLOH event taken place. If yes, would the genetic distance be close if it was discounted? If yes, still consider possibly related.
    3. b) If the genetic distance is large, have the two got a "group" of distinctive off modal slow mutating markers in common? If yes still consider possibly related and investigate further.
    4. c) If the genetic distance is large, is there a large difference on a single marker? On rare occasions a multistep mutation can occur on a single marker. Single mutations of 4 steps appear to be possible, and perhaps even rarely up to 10 steps. If two match on 66 markers out of 67, but have a 10 step difference on 1 marker, consider possibly related. But if two have multi step differences on 3 or 4 markers, it is not likely they are all single mutations of multi steps, so consider less likely to be related.
    5. d) Be sure to count the genetic distance on DYS389 correctly. Comparing 10,26 to 13,29, is a 3 mutation step difference, not 6 steps difference, and in fact could possibly be only a single mutation of 3 steps.
  3. 3) If the genetic distance is small, do the two have the same surname, or a possible variant of the same surname. If the same surname, consider possibly related.
  4. 4) If the genetic distance is small, do the two families trace to the same village in the same time period? If yes, consider possibly related.
  5. 5) If the genetic distance is small, do the two share any distinctive off modal slow mutating markers. If yes, consider possibly related.
  6. 6) If the genetic distance is small, are they both close to a very common haplogroup modal? If both are an exact match for a very common haplogroup modal, consider that the two may be unrelated unless other evidence suggests a link between the two.
  7. Once you have been through these series of preliminary screens, you should be down to a short list of "likely related", and a moderate short list of "still possibly related". You can then massage the short lists to refine probabilities. Some things I do are...
  8. 7) If the haplotypes have not been SNP tested, look for close matches on Y-Search. If there are very close matches on Y-Search, have any of these close matches been SNP tested? If they have, you may be able to establish that your two prospective haplotypes are likely from different haplogroup clades. If so, they are likely unrelated.
  9. 8) Do any of the close matches of different surnames on Y-Search have long known pedigrees. My closest non surname matches on Y-Search are 60/67, and 61/67, both of the same surname to each other, and one has a continuous paper trail pedigree back to the year 1100. FTDNA TiP estimates I have around 95% chance of a common ancestor with these two within 24 generations, say since the year 1200. If TiP were right, then it is possible I descend from the same named ancestor living in 1100 as these two matches.
  10. 9) If you have two persons of a surname you are comparing, what is the possible origin of the surname. If the name is Smith, Taylor, it may not give any leads. But if the surname is a rare surname, supposed to descend from a known founder say in Norman times, then research that founder. If it is a well known prominent person, he may have left documented male lines to the present. Search on genealogical sites for male line pedigrees to the present from that person. Sometimes the surname changes in the male line, so the surname may be different in the present. If you find any male lines to the present, check to see if they have already been DNA tested. If not, try contacting known descendants to ask them if they would consider DNA testing. If a purported surname founder is similar to the haplotypes you are comparing, it might support them being linked.
  11. 10) In searching for potential origin of a surname, focus on the individual area where the haplotypes came from if known. My surname had many different origins in different locations in the UK and Europe. My Y line only traces back to 1789 in a specific UK village, but Y-DNA tested people of my surname from the same village match my Y-DNA, and trace back to 1560 in that village. Further, I have with research established that there was only one documented family of that surname in the village in 1524 who according to wills only had sons, and that family was likely descended from a person of the same surname in the village in about 1450. The Y-DNA line may therefore have been under my family surname as far back as 1450 in my ancestral village. I investigated if my surname had been in the village earlier. I found that an earlier spelling of my surname had been in the village around 1360 when the senior line of the family ran out of male heirs and the principle family estates were passed on via heiress daughters. At that time, sometimes surnames were passed on by daughter heiresses if there were no male heirs. Looking further back, an even earlier variant of my surname first appeared in the village in about 1150. Research indicates that the person in 1150 was almost certainly from a well documented Norman family, and from land deeds, it is possible the land held by the same surname in 1360 was the same land or nearby land to that held in 1150. Further research indicated that a documented male line from that Norman family came down to the present, was DNA tested (when I asked them to test), and was from a different haplogroup to me. So there is a building hypothesis to test... A family of my surname was in my ancestral village apparently continuously from 1150 to 1360, when it was documented the senior line at least of the family ran out of male heirs. Persons "of my surname" in the village possible carried my Y-DNA back to about 1450. A family of different surname to me with a pedigree back to 1100 is a 61/67 match to me, and matches 5 of my 6 distinct slow mutating off modal markers, and was in the general neighbourhood of my ancestral village between 1220 and 1450. FTDNA Tip estimate I have around 95% chance of a common ancestor with that different surname family since about 1200. This suggests a time period of opportunity for a connection to the different surname.
  12. 11) By DNA testing males of my surname in increasing sized circles from my ancestral village I have found half a dozen others who match my Y-DNA.
    Every one of those was a person who I tracked down from genealogical web sites, and telephone books etc, and asked if they would test. All of these lines except one may have originated from or near my ancestral village in around the 1600s, but one from London could hypothetically be from my ancestral village in the 1700s, or perhaps be from a deeper source in his London area in the 1300s, or from a nearby area to my ancestral village in the 1500s.

I am waffling on a bit, but I guess I am saying you go through a hierarchical series of screenings, and as the list of probables shortens, you do some hard work to try and put things into a context. Hopefully this will lead you to some sort of hypothesis which can be tested by DNA or other means.

By measuring the height of a boy, you might be doing a first screening to reduce the field of probabilities for which grade he is in. But that first screening needs to be followed up by successive screenings of more information before you can develop a high degree of certainty as to which grade he is in.