No announcement yet.

Establishing surname - # Matches or Genetic Distance as priority?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Establishing surname - # Matches or Genetic Distance as priority?

    I'm hoping to establish my great grandfather's surname. When I look at my Y-DNA (37) matches there are none with an obvious link to my tree.

    I have 4 with the same surname (McPherson) with a genetic distance of 1-2. But then I have a second surname (Campbell) with a much higher count of total matches (12), but a higher number matches. The second surname fits with some of my paper based research. If I do a pareto of the # matches, Campbell is the top of the list.

    Could someone who understand better than me advise the best way of interpreting this information?

    Surname Count of Last Name
    Campbell 12
    Mitchell 9
    Young 7
    McDonald 7
    Sutherland 6
    McPherson 6
    McCoy 5
    Stewart 5
    Morrison 5
    Forbes 5

  • #2
    Run TiP Reports on those matches, it will look at the GD and the mutation rates for those markers and make predictions from there.

    Surnames(that weren't Clan names and/or linked to nobility) only really started becoming common in much or Europe circa 17th Century, Assuming a 20 year generation (5/century) you'd be looking at surnames with consistent high probabilities on the TiP which are closer than 20 generations back(If you go with the 33 years/male line generation, the window narrows to as little as 12 generations).... And preferably you would do so on a 67 or 111 marker STR Test, 37 marker results can be rather misleading, but they're better than nothing.

    If the TiP report says they're not likely to be related within the 12 to 20 generation range, they may not be the surname you're looking for(a higher level test might change the results, but until that happens, work with what you have).


    • #3
      Thanks for this. If I look at my, say, top 5 "Campbell" results, there's an over 97% probability of being related in the last 12 generations. This drops to 58% by 4 generations with all but 1 match, where that probability stays up at 83.49% at 4, 97.28% at 8.

      When I look at my top 3 overall matches (all McPherson), there is a 83.49% probably at 4 generations, rising to 97.28% by 8 generations.

      I'm now struggling to get my head around this. Both scenarios can't be right - or can they? I may be missing something obvious (likely, as I have no real idea what I'm doing yet!).

      Thanks again


      • #4
        The history of surnames is complicated! There was often a period when surnames were in use, but they weren't yet stable and transmitted by paternal descent, as we often understand them today. In many areas, surnames didn't become completely stable until perhaps the 16th Century or even later -- not to mention areas where patronymics or other systems were in use until quite recently. You can regard the matches as "maybe", but the answer could easily turn out to be something else entirely. Don't rule anything out at this point, apart from individual families that definitely don't match your Y DNA profile. Consider Big Y, if you really want to be able to find out where in the Y haplotree you belong.


        • #5
          My Haplogroup admins caution against getting too excited about 37str matches and jumping to conclusions with them. It is part of why I said testing at level higher than 37 markers is preferred. You could see that high probability GD2 match at 37 markers turn into a single digit probability at 67 or 111 markers, assuming they even appear on your match list at that level.

          It is entirely possible several of those 37 marker matches have a MRCA on the Y-DNA line more than 2,000 years ago after SNP testing. 37 markers is where the results start to become "genealogicaly useful" and even then it has a margin of error. (I'm GD2 with my own father @37 markers, but as I share 3384cm with him on the FF atDNA test, it's highly unlikely he's not my dad)

          The more markers tested on which you match with few to no significant differences(specifically on low mutation rate markers--which the TiP report is looking for) is important. It is extremely reasonable to expect someone related to you within a handful of generations on the Male line will show within Y37's filter criteria. But that doesn't mean everyone who meets the filter criteria is closely related. As you work up the line, the probability of a match being related or not will shift as well. You can see this by doing TiP reports on your Y12 matches at Y12, Y25 and Y37, watch how the numbers shift around.

          As it sounds like you have a good cross section of people at Y37, chances are you probably do have a (somewhat) close relative present, so one of those surnames may be the right one, or it could be none of them are. If any of them show at Y111 at a close GD with a good TiP, I'd be somewhat surprised if it isn't the right one. But even at Y67, you can find distant non-surname matches. My only Y67 match outside immediate family doesn't share my surname and appears to have been in the US over 100 years before my patrilinieal line arrived for example. Although based on Y67, TiP thinks we're very likely to be related within 20 Generations(400 to 700 years), that still places us outside most surnam practices being useful.