Herriott Surname DNA Project
A 5,000-Year Tree for (Most of) Our Family Lines
By Scott R. Herriott
In the previous article (December, 2016), we reported the results of an attempt to use DNA data to show, in a single family tree, the relationships among all the Heriots, Herriots, Herriotts, Harriots, and Harriotts whose male-line DNA is clearly Scots. This article reports our continued work on that matter, leading to an answer to the question about how our "New Jersey" group, descendants of the immigrant David Herriot (ca. 1660-1725, containing spelling variations Herriott, Harriot, and Harriott) is related to the Heriot line associated with the founder of the Heriot School in Edinburgh, George Heriot (1563-1624).
Two Types of DNA Analysis: STR and SNP
DNA analysis gives us a perspective on family relationships that goes farther back in time than historical and genealogical analysis. If we say that genealogy is like looking into the past with our bare eyes, then DNA analysis is like looking with binoculars or a telescope. In fact, there are two types of DNA analysis, one for the closer past (binoculars) and another for the more distant past (telescope).
Since the start of our DNA project in 2007, we have been using the binocular form, which geneticists call STR (single tandem repeat) analysis. In the past few months, we began using the telescopic form, called SNP analysis (single nucleodide polymorphism). A word about what DNA is will help us understand the difference between these two methods.
The human DNA is a molecule that looks like a chain with 3 billion links, called nucleotides. There are only four types of nucleotide, known by their initial letters A, T, C, G. These are the four "letters" of the DNA code. The STR and SNP analysis refer to two different types of change in the DNA molecule. In the STR, a small segment of the code gets repeated some number of times at a particular location on the molecule. For example, the segment ACTC might appear several times in a row, and an STR-type change in the DNA would be to add or subtract one complete segment. Thus, STR tests are reported as the number of counts of the segment at a particular location on the human DNA molecule. In contrast, an SNP ("Snip") change is the change of one letter only, at a particular location.
The STR changes give what I call a binocular view, because they allow us to look back in time, with some reliability, between 500 and 5,000 years. STR changes take place more frequently than SNP changes. The STR tests that we have used since 2007 in our DNA reports from Family Tree DNA included 37 or 67 locations ("sites" or "markers") on the male (Y-) chromosome of the DNA molecule. Those locations vary from some that are relatively "hot", meaning that a random change in the STR count takes place more than once every 150 generations (>0.7% chance per generation) to "warm": (1 every 150-300 generations, 0.3-0.7% chance per generation) to cool (1 every 300-1,000 generations, 0.1-0.3% chance) to cold (less than 1 per 1,000 generations, so <0.1% chance per generation). The hotter sites give us a clear look only into the near past, because an STR change can be reversed and thus obliterated from view by a subsequent change increasing or decreasing the count at that site. The cooler sites give us a view into the deeper past, those changes stay in the DNA for many more generations but even they are not reliable beyond about 5,000 years into the past.
The 60,000-Year Human Family Tree
In contrast to the STR changes, the SNP ("Snip") changes give us a telescopic view of our past. SNPs take place at sites on the DNA molecule that are so "cold" that, once they occur, they are almost never reversed. So SNP changes get carried all the way down the male line of descent. As a result, they show changes that took place from about 1,000 years ago to nearly 100,000 years ago.
Geneticists use SNP changes to construct a family tree for all humans, starting from the humans who left Africa 60,000 to 100,000 years ago. Here is the last 60,000 years of that tree.
Geneticists used combinations of letters and numbers to denote the branches of the human tree created by changes at specific SNP sites. For example, most males whose families come from Western Europe have a SNP called "R1b", shown in the lower-left of the graphic. Of our 10 or 12 distinct "Herriott" lines discovered in our DNA project, 9 are forms of R1b. However, 3 lines, including our "Belfast" group of Herriott/Heriot/Herriot/Harrietts and our "Berwickshire-Upon-Tweed" group of Herriott/Heriot/Harriotts, are from the I2 and I1 lines, respectively (center of the graphic), which are Scandinavian.
The 5,000-Year Herriott Family Tree
Using both STR and SNP test results, we can now look back at the last 5,000 years in our R1b lines. Until this year, we were using only the STR data, which can be confusing due to the possibility of the same STR change taking place independently down two lines of descent (creating a false appearance of near-term ancestry) and due to the possibility of a change being reversed down one line of descent (false appearance of distant ancestry).
The results of our recent SNP tests have partly confirmed but also changed our understanding of how even the various R1b "Herriott" lines are related. Answering one longstanding question, we found that the Heriot line associated with the jeweler George Heriot has SNP type "P312", whereas the New Jersey line has SNP type "U106". Their relationship is seen in the SNP tree, which shows both the new SNP marker name (e.g. L23) set by the ISOGG in and the previous name (e.g. R1b1a1a2a) used by Family Tree DNA. The P312 and U106 family lines have a common descent through markers L23 (about 6200 years ago), L51, and L11/P310, but the P312 and U106 lines diverged about 5,000 years ago. The P312 line descends through a marker named L21 to the Longniddry Heriots. L21 is also known as the "Scots Modal" genetic group or haplotype. Our New Jersey group, being U106 (4800 years old), is not within the Scots Modal, as we had assumed when creating the trees published in the December 2016 article.
SNP Analysis of the Other “Herriott” Family Lines
We have just started doing the SNP analyses for lines other than the Longniddry Heriots and the New Jersey group. One of our DNA participants, a member of the East Lothian group of Herriots, had paid for small (and cheap) SNP test. They turned out to be within the L21 Scots Modal group but descended from a branch (called DF49) that separated from the Longniddry Heriot group about 4,300 years ago. For the Longniddry Heriots, we did the most comprehensive DNA test available, called the Big Y test. It traced their line through L21 to marker DF13 (4300 years old) to marker CTS1751 (4200 years old.) That was the most recent SNP marker, contained in the Longniddry DNA report, that has been identified by genetic genealogists. We may learn more in the future, because new markers are discovered every year as more people contribute DNA data for comparative analysis.
The STR analyses that we reported in the December 2016 article showed that the Berwickshire Herriots seemed to be rather similar in DNA profiles to the East Lothian group. We will soon proceed with SNP tests on the Berwickshire group to see how closely they are related to our known East Lothian DF49s.
Our STR analysis also seemed to indicate that two of our ungrouped participants, Michael Herriott and and John Herriott of the UK, seemed to be more closely related to the New Jersey group than to any other group in our project. We ordered a test to see if they are U106. John's results came to us just before this article was published. He was not U106, so we are left guessing how he is related to the Longniddry and Berwickshire groups. Michael's U106 test is still pending.
The U106 is an interesting SNP marker, because it is also revealed through the STR 67 marker tests that we have been using for the past ten years. There is a very cold STR marker (called DYS492, averaging 1 change per 5000 generations, 0.02% chance per generation) that Family Tree DNA has in its 67-site Y-DNA STR test. This STR marker correlates very highly (95%) with the U106 SNP marker. A count of 13 repeats at DYS492 indicates the U106 haplogroup. A count of 12 repeats indicates an R1b branch other than U106. We used 37-marker tests for most of our early DNA studies, but we can upgrade the test to 67 markers if needed. In the next article, we will report on how the 67-marker STR tests give (or not) a more close-up view of the lines of descent within and around U106.
Our "New York" group of Harriott/Harriets is not obviously close to either the P312 Longniddry Heriot line nor the U106 New Jersey line, so it's hard to know where to start with SNP testing on them. The most comprehensive SNP test (the Big-Y test from Family Tree DNA), which gets data on 25,000 SNP markers, costs $575 per person. Cheaper packages, like the U106 test, cost about $169, but they are useful only when we are fairly certain that a person is in the U106 group.
Our Herritt participant in the DNA study, though identified in our old STR test as an R1b, is so different from our other lines that we would probably have to use the Big-Y test to learn anything more about his place on the SNP tree. Our Harrod participant seemed to have an STR test not too far distant from the Berwickshire group, so if we can get an SNP result for Berwickshire, we may be able to test Harrod inexpensively.
In the next article, when we have more SNP test results, we will publish a tree showing how our various lines are related in SNP terms.
Who are the U106 "New Jersey" Herriotts, if not Scots Modal?
The U106 genetic line is not the primary Scots group, but it has its own distinctiveness. Members of U106 include the discoverer of DNA, Dr. James Watson, and the other scientist who along with him was the first to have their DNA mapped, Dr. Craig Venter, as well as a notable advocate for genetic genealogy, Dr. Spencer Wells.
Wikipedia's article on U106 gives another angle on the membership:
R1b-U106 is a patrilineal descended family that appears to descend from an ancestral R1b group located among or near the Yamnaya culture, north of the Black Sea area. The group rose to significance in southern Germany and the surrounding areas about 3000 BC. Although U106 is found all over Europe, and in countries that Europeans have migrated to, it is most significant in Germany and surrounding countries, Scandinavia, and Britain. Depending on which branch of U106 a member descends from, the people on that branch adapted to a variety of different cultures along the way, including various derivatives of Slavic, Latin, Celtic, Belgae, Saxon, Viking, and other cultural groups. U106 is a family, not a culture. Some families of the historic nobility have paperwork ancestry that reaches back farther into The Genealogical Gap of the Dark Ages. This enables us to compare the paperwork to the DNA as follows: both the Bourbon family of the Spanish and the former French Royal families, and the Wettin family of Saxe-Coburg from which the British and the Belgian Royal families, and the former Portuguese and Bulgarian Royal families descend, are confirmed to descend from the Z381 Branch of R1b-U106 [as is our New Jersey Herriott group]. The Bourbons are highly probable to descend via Z156, while the Wettins definitely descend via Z156 and its subclade DF98.
Now you can see why the Family Tree DNA web page about U106 refers to us as the Haplogroup of Scientists and Kings.