Statistics and Significance

[rbsvnski]

Is there some place on FTDNA or elsewhere that can shed some light on the significance of our test results? My 12 marker YDNA resulted in no matches. I have read in the forum about individuals with over 100 matches in the data base. I would be very interested in knowing the mean/median/mode number of matches that tests have yielded to date. How unusual/common (percentage-wise) is zero matches? I could come up with some other useful statistical analyses but you get the idea.

Since alleles have little inherent meaning until they are compared to some larger database it seems that other ways (beyond simple no. of matches) of analyzing/comparing one's results to the whole would be useful and possible.

[josh w.]

It is not at all unusual to have zero matches. It depends on your haplogroup and country of origin as well as the national location of the testing firm. Ftdna is located in the U.S. where most European immigrants come from northwestern Europe. People with large numbers of matches tend to have this background. It is possible that you have a very unique haplotype but it is more likely that you will have matches as more people are tested. It is also possible that you will find matches at other databases such as Y-hrd or Sorenson. Ftdna does indicate what percentage of its total sample shares a particular haplogroup. The information is listed under the "statistics" folder at Y search. It is also possible that you might find matches by uploading your markers to Y search but this database greatly overlaps the general Ftdna database.

[cacio]

rbsvnski:

you have no 12/12 matches, but you have a couple of distance 1, and several more with distance 2, so probably you an already infer something about the distribution of your distant ancestors. As josh was saying, outside of the typically British/Irish R1b and I1a, matches are very rare. Among my (uncommon) L, nobody has any match, and typically the closest neighbor has a distance of at least 2 or 3. Until thousands of people with your own geographical origin test, this is bound to be the case.

If you want to know more about your lineage, you could do more tests. First, you could test more STRs. If you have no match at 12, you won't at 25, but distances become more informative. For instance, a distance of 1 at 25 markers may mean close relation, something not necessarily true at 12 markers. Second, you could do a SNP deep clade test, to establish which subgroup of I you belong to. Both FTDNA and Ethnoancestry offer deep clade I tests. This won't give you matches, but will give you information about the distribution of your group.

cacio

[lgmayka]

Is there some place on FTDNA or elsewhere that can shed some light on the significance of our test results? My 12 marker YDNA resulted in no matches.
...
Since alleles have little inherent meaning until they are compared to some larger database it seems that other ways (beyond simple no. of matches) of analyzing/comparing one's results to the whole would be useful and possible.

1) Right now, we aren't even absolutely sure you're in haplogroup I. As you know, FTDNA has not SNP-confirmed the haplogroup of anyone within 1 step (genetic distance) from you. (Some 2-off matches are confirmed as I1a or simply I.)

Your deepSNP test was due on August 28. I suggest you ask Eileen Krause (mailto:[email protected]) when you can expect your results.

2) Among your 1-off matches on Ysearch is a Ukrainian, CAD25. Offhand, I would suspect that he is your nearest match right now. I suggest that you send him a message (through the Ysearch mechanism) pointing out your near-match and suggesting (a) that he and you both upgrade to more markers to test the match, and (b) that he join our Polish-Lithuanian-Ukrainian-Belarusian-Latvian project.

3) Most Slavs do not get exact matches even at 12 markers, except sometimes in the majority haplogroup R1a. What we do instead is use Ysearch to find those who are nearest to us, make hypotheses, and look for evidence.

I myself still have absolutely no exact matches at 12 markers. But I have upgraded all the way to 67 markers. Here is a list of my near neighbors on Ysearch. This list includes anyone who is no more than 7 steps away at 37 markers, or no more than 37 steps away at 67 markers:

http://www.ysearch.org/search_result...ting_marker=30

Unfortunately, I am probably not close enough to any of these men to establish a genealogical connection to them. However, our 'neighborhood' does demonstrate my Polish Carpathian ancestry. This neighbor list also shows something extremely peculiar: There is absolutely no one in Ysearch who, at 67 markers, is between 16-37 steps away from me. This seems to imply that Dinaric (Slavic) I1b forms a tight coherent block totally separate from everyone else (including the other varieties of I1b). In fact, with the exception of the 4 men you see listed close to me, my nearest 67-marker match is a genetic distance of 50 away, and he claims to be of haplogroup D2!

http://www.ysearch.org/search_result...ting_marker=13

4) If you are seriously interested in making genealogical connections, the first step is to invite men with your surname (or variations of it) to be tested. You are lucky in that your Polish surname is typically nobility (-ski), not serfdom, and so it is probably older and more likely to have resulted in a bushy but connected family tree (as opposed to a serf name like mine, which is probably only a couple of hundred years old and may have been 'invented' in several different places independently).

Of course, once you convince others with your surname to be tested, if they match you, you will then have to convince them to upgrade to more markers.


So I must again suggest that you upgrade your marker count whenever you can afford to do so. I cannot guarantee that you will be able to make a genealogical connection, at least not immediately, but you will at least learn more about Slavic population history!