The DENNISON Singleton Patrilineages DNA Project

This project is in no way affiliated with the DNA testing company, Family Tree DNA or any of its surname projects. Rather, the members whose ancestral and DNA data are presented here are members of the independent DENNISON Surname Patrilineages Association (DenSPA).

Navigating from here

The menu buttons at top right take you to other pages on this site, while the nav panel above targets other points on this page, or brings up other resources (papers I’ve written, and the like). If you find yourself lost, the browser BACK button will take you back to where you were (some people also have a convenient BACK button on their mouse, right under their thumb). Or hitting the HOME key of your keyboard will take you back to the top of this page where you are now.

Some Key Terms: haplotype, haplogroup, patrilineage, RPH.

DENNISON Singleton Directory of Researchers

Active researchers of this DENNISON patrilineage are shown below. Those with highlighted names may be e-mailed by clicking on their names, and their posted descendancies may be viewed by clicking on their highlighted Patrilineage Project#s. Satellite members of the project are listed immediately after the principal researcher for the sublineage they are interested in, and their names are preceded by a dash. Where the person tested is not also the principal researcher, the former’s name appears under the latter’s, in parentheses.

DENNISON Patrilineages and the R1b Haplogroup

Executive summary: At present all the DENNISON singleton patrilineages belong to the haplogroup R1b, though by definition no two of them are related within genealogical time. This analysis, based largely on the haplotype comparison matrices below, is an attempt to quantify very roughly just how unrelated these patrilineages are, i.e. how far back their common ancestors go. The answer is: mostly, the ancestor common to any particular pair of patrilineages lived sometime between 700 BC and 300 AD.

The one exception is the ncestor of Kathy-03: he was probably a descendant of the Irish chiefly line of the Ui Neill, and may have walked the earth as late as the period 800-900 AD.

Collected on this page are all the DENNISON patrilineages with only one tested member. At present, all of these, and also the two principal patrilineages, DENNISON Patrilineage 1 and 2, all belong to the R1b haplogroup, which is the predominant haplogroup in northwestern Europe, and especially in the British Isles. As a result of this, no haplogroup has been more extensively parsed into subclades and sub-subclades than R1b, and as more recent ySNP mutations continue to accumulate, we can hope that in the not too distant future the ySNP mutation tree that sorts human population lineages into ever more articulated branches will, besides providing a fascinating companion to our archaeological knowledge of ancient cultures and their history, rise to the level of genealogical significance.

In the meantime, some may find it of interest to consider all the DENNISON patrilineags together as a sample of R1b in Britain. If nothing else, it should provide some perspective on where things stand at present with this type of analysis, and prepare the mind for the discoveries to come.

Accordingly, I have constructed the usual brace of haplotype comparison matrices to encompass all the DENNISON patrilineages: one showing, pairwise, their relative Genetic Distance (GD), and the other their pairwise TMRCAs.

Haplogroup R1b in Europe, Britain, & elsewhere

R1b is the most common haplogroup in western Europe, and it is overwhelmingly prevalent in the British Isles, as shown by this distribution and concentrations of R1b, and farther down the same page is a chart of estimates for the age of each of it’s principal SNP branch points. Most Brits belong to the broad haplogroup subclade R-M269 (defined by ySNP mutation M269) and so do, at present, all ySTR-tested DENNISONs.

Where the DENNISONs Fit into Haplogroup R1b

Thanks to the additional ySNP testing by Patrilineage 1 member Alan-05 we also know that Patrilineage 1 falls into haplogroup R-U106* (which branched off the main trunk of R1b about 5000 years ago), and additional testing by Lee-11 shows that Patrilineage 2 belongs to R-L21+ (which branched off about 4000 years ago). Kathy-03 (a cousin of Terry-16 who has deep clade tested as R-M222, branched off of L21 about 2000 years ago, and is closely associated with the old Gaelic cultures of Great Britain, and particularly with the ancient chiefly Ui Neill clan.

“Niall of the Nine Hostages” DENNISONs

R-M222 is a species of what was originally originally called the Irish Modal Haplotype (IMH) that was thought to have characterized the line of Ui Neill chieftains, of whom the most famous was the semi-mythical 5th century figure, “Niall of the Nine Hostages”. While there probably was such a man, just as there was probably a King Arthur (Niall’s rough contemporary), in view of what we actually know about these men and their exploits, based mostly on many centuries of storytelling, it may be too much to call them historical figures. However, we do know that there was a line of prominent northern Irish chieftains called the Ui Neill, who formed an hereditary patrilineal ruling clan who are believed to have practiced extreme forms of polygyny, and thus to have collectively fathered a large proportion of their own tribe.

And we also know that the haplotypes of an extraordinarily large proportion (from 5-20%) of those who trace their ancestries back to an area of northern Ireland centering on County Donegal, but reaching out also to western lowland Scotland appear to converge on a common male ancestor who lived roughly 1500 years ago. By far the best hypothesis to account for both phenomena, is to suppose these folks are all descended from this chiefly Ui Neill clan.

I should note that the inadequate “bikini” IMHaplotype of just 17 markers has since been hypothetically extrapolated to full 111-marker size—sufficient, perhaps, to justify a TRMCA estimate for the whole group which is not inconsistent with this descent hypothesis. However, it should also be understood that: (1) the only definitive marker for this subclade is the M222 SNP mutation, and not all R1b’s with this marker need have descended from the Ui Neill chieftains; but on the other hand (2) the MRCA of any particular pair, or set, of descendants of the Ui Neills is likely to have lived many centuries after Niall of the Nine Hostages. Thus, to reliably classify a haplotype as belonging to the Ui Neill descendancy, we should expect to find TMRCAs of less, or even substantially less than the 1500 years or so back to Niall himself, or to whomever was the founder of this proliferant polygynous tribal aristocracy.

There has been intense interest in this particular Irish and Scottish descent group, with many projects and studies devoted to it, not all of them DNA-related. The most comprehensive of these is the FTDNA R-M222 project, and this collection of links relevant to R-M222 is the best I’ve been able to find.

The GD and TMRCA charts I’ve constructed are misleading in one respect. Although I have included the R-M222 modal value to provide some notion of how closely each other tested haplotype resembles it, the numbers in the cells of these matrices are supposed to represent the GD or the TMRCA between pairs of tested modern descendants, whose haplotypes have both had opportunities to pick up mutations since they began to diverge from their mutual MRCA. Thus, the only estimate that’s relevant for classifying Kathy in the Ui Neill group is the estimated 1280 years back to a common ancestor—one who was born, probably, no earlier than 650 AD, and who therefore probably fits within the Ui Neill descendancy. Because I have used a stringent confidence factor of 90% in creating this TMRCA chart, there is only a 10% chance that the common ancestor goes back farther, and in fact it is probable that the common ancestor lived several centuries later than the 7th. It is even theoretically possible, given that some of these old Irish lines adopted surnames as early as the 900s, that the common ancestor of Kathy, and Terry-16 of Patrilineage 5, was surnamed DENNISON, however the proliferation of rather close matches to Kathy’s haplotype makes it overwhelmingly likely that DENNISON was adopted rather recently in her line: probably within the last 300 years.

The addition of the M222 modal value, though, provides some guidance as to the distribution of mutations in each pair of lines, and that is a bit anomalous in this case. Terry is shown as have a GD of 10 from the M222 modal, while Kathy’s GD is only 3. One of the M222 links referenced above provides an extended spreadsheet of GDs, by surname, and the mean and modal GD appear to be about 5, with the maximum GD begin 13, for an M222-confirmed member of the project. Thus our pair of DENNISON members fit comfortable within the normal range. I might note that neither Terry’s nor Kathy’s haplotypes are included in this or other R-M222 project databases, because neither appear to have joined that project.

Just How Closely Are All These DENNISON Patrilineages Related?
The Pitfalls of TMRCA Estimates

The TMRCA chart I've constructed to show the degree of inter-relatedness of all DENNISON patrilineages is highly misleading in several respects, and it is different in one way from the TMRCA charts for individual patrilineages. Let me take up this latter point first.

The TMRCA charts for DENNISON Patrilineages 1 and 2 show the expected number of years back to the common ancestor for each pair of haplotypes represented in the chart. The actual number of years back to the common ancestor is unknown, and based on the probability factors we have to work with can cover a wide range of many hundreds of years, and the set of possible values could be plotted on a graph as a frequency distribution. In the terminology of probability and statistics, the expected value for such a frequency distribution of possible values, is the mean (or average) of all the distributed values. Thus the number in each cell of the TMRCA matrix represents an estimate of the number of years back to the common ancestor of one pair of haplotypes, such that there is an equal probability that the real number is less than, or greater than the expected value.

Unfortunately, the expected value tells us nothing about how widely distributed the possible values are around this mean, and in the case of TMRCA estimates, the answer is: very widely distibuted indeed, with an especially extensive skew back into the past. Thus, with a TMRCA estimate of say 1000 years, there is quite a high probability that the real TMRCA might be as little as 700 years ago, and an equally high probability that it might be 1500 years ago (skew means that the distribution of possible and probably real values is asymmetrical).

In fact, the range for an expected value chart is so wide as to make TMRCA of little use for genealogical purposes, but it can help to place common ancestors roughly within the broad matrix of history, and pre-history, with a little tightening of the distribution range. Tightening can be accomplished by calculating the TMRCAs as upper boundary conditions, and that is what I have done for the all-DENNISON TMRCA chart. Instead of using a 50% probability level to produce an expected value chsrt, I have used a 90% level. This means that the number in each cell represents the earliest a common ancestor may have lived, to a 90% confidence level.

I created the chart this way because I wanted to show, if possible, that the lifetime of Kathy and Terry’s common ancestor fits within the historical framework of the dynasty of Ui Neill chieftains, and that is does (although even to, the TMRCA numbers in the chart are slight underestimates, because they do not take account either of possible back mutations, or of the fact that we can say with confidence from the genealogical evidence that these two aren’t related back for a certain number of generations). Still, we have direct evidence from SNP testing that Terry belongs to this line, and Kathy’s close match to the Ui Neill modal places her in incontrovertibly into the same group. Thus, the TMRCA estimates are merely confirmatory.

The GD chart provides a better measure of relatedness than TMRCA, though it is a measure of relative relatedness, not an absolute relatedness calibrated to a time frame. However, this Ui Neill descent for two of our DENNISONs provides some time calibration. The GD of Kathy and Terry with respect to each other is 13, and we can pretty much assume from what is known about Irish history that this places the common ancestor no farther back than about 500 AD, and more likely a century or two later. Using this GD=13 number (roughly a count of the total number of mutations accumulated in both these haplotypes) and a different calculator than the one used in constructing the TMRCA chart (and assuming 32 years per generation, rather than the 34 I normally use), the expected time of birth of their common ancestor can be calculated to be about 800 AD, and factoring in those other variables, might bring that number down to, say 600 AD, which, again, fits with the history.

If GD=13 carries us back to about 600 AD, how far back does the GD=16 which separates Patrilineage 1 from Patrilineage 2 take us? That happens to be a less straightforward question than meets the eye, because the haplotypes I’ve used to represent these patrilineages, are their respective RPHs, which are meant to represent the haplotype of the common ancestor of each set of patrilineage haplotypes within genealogical time. We thus need to project the additional GD of 16 back from the time of the common ancestors of Pat 1 and 2, in order to arrive at an estimate of the common ancestor for all the members of both of these patrilineages. The highest GD found in the 37-marker chart for Patrilineage 1 is 8, and for Patrilineage 2 it is 9. I thus estimate that the GD between these patrilineages is 16 + 8 = 24, and running that through the TMRCA calculator under the same assumptions as above, and making roughly proportionate adjustments for back mutations and genealogical knowledge, I get a TMRCA to the common ancestor of both patrilineages of about 450 BC. Or, using the time calibration, by which GD=13 projects back to 600 AD, a GD of 24 would project back to 550 BC.

So, it appears that the common ancestor of DENNISON Patrilineages 1 and 2, lived about 500 BC.

The most striking feature of the GD matrix is how close Denny-18’s haplotype is to that of Patrilineage 1—the difference between them is just GD=12. However, this number needs to be adjusted to take account of the fact that the Patrilineage 1 haplotype used is that of the RPH and therefore has no more than about 1 mutation, while, as the Pat 1 GD matrix shows, the typical descendant of that lineage has picked up about 4 mutations. Thus, the total GD for Denny is probably about 16, and his common ancestor with the Patrilineage 1 folk, would project to about 300 AD.

I would guess that these are probably pretty typical numbers for the inter-relationship of two Brits, selected at random.