THE DNA PROJECT:
If you've been here before, and are just trying to find the latest update of our results, click on one of the following:
My name is Fred Coffey, and I am the "Web Master" of the Coffey Surname Project web page. If you have any questions on the following, please feel free to email me, or our co-administrator Lorie Okel:
If you are male and your surname is Coffee or Coffey or anything that sounds like that (maybe even "O'Cobhthaigh" if you're still using one of the original Gaelic spellings?) you're invited to join our project. Read below to see what we know so far, and consider if you might "fit in".
You can join our DNA Project by visiting www.ftdna.com, inserting “Coffey” in the name search box, clicking the "Coffey" project, and filling out the form.
NOTE: There are other DNA testing services, such as www.ancestry.com , and we invite those tested elsewhere to get in touch with FredCoffey@aol.com . There is enough overlap that we can probably determine if you have a match with anyone in our project. We have, for example, identified several “ancestry” tested individuals who match our Edward Group, our Northern Group, and our County Meath Group (see below), and a few of them have joined our project and are included in the genealogy page.
WHO ARE WE?
The Coffey name (using this one spelling for convenience) is of Irish origins. Our belief is that there were several different and genetically unrelated groups of Coffey's in Ireland, and our DNA studies indeed do confirm this. Does DNA offer any clues about connections to specific Irish Counties, and Coffey Clans? For some preliminary thoughts after you read the discussion below, visit
Diversity: There are presently 26 Coffey "y-lines" identified (including only those who have agreed to publish their info). We define these as tested groups or individuals who have the Coffey surname but who have a y-DNA profile different from all others, OR those who have known Coffey-connected y-DNA, but different surnames.
And for those with the Coffey surname, we know some had a DNA origin in Ireland, some in America (we have stories for most of those), and many "unknown". They break down as follows:
Coffey Surname/DNA Origin Ireland 9
Coffey Surname/DNA Origin America 4
Coffey Surname/DNA Origin Unknown 6
Other Surname with Coffey y-DNA 7
We've started identifying these lines/groups by various names (usually the oldest ancestor in each Group, based on genealogy research of some of the tested people, or by geographic origin):
Edward Group: Edward Coffey, who married Ann Powell, is believed to have come to America in the 1600's. So far we have about 37 tested people with Coffey/ee surname who are genetically related to Edward, our largest group. (We are now beginning to see some minor genetic differences, and believe we are beginning to identify some sub groups.)
Peter Group: Peter Coffee Sr. was born in Ireland in about 1716. We have now had three people tested who trace their lineage back to Peter. However that test proved something that had only previously been suspected - - Edward and Peter were related! Since genealogical research proves that Peter is not descended from Edward, the Most Recent Common Ancestor (MRCA) has to be farther back in Ireland.
Keogh/Kaho/Kehoe Group: We now believe Edward and Peter were part of a much larger family, involving, in addition to Coffey, the names Keogh, Kaho and Kehoe. And we believe the combined family Irish homeland might be in the area of Wexford, Kilkenny and Tipperary Counties. There’s a lengthy discussion at:
This “Kehoe” groups’ data, by the way, appears only on our “DNA Data” sheet, and NOT on our “Genealogy” sheet. We don’t know the genealogical connection. DNA proves the connection to Edward and Peter is absolutely real, but it has to be back in Ireland a few hundred years ago. The “paper trail” for our genealogy doesn’t make it back that far!
"Edward Connections" Group: The "Edward" Group might be considered even larger than the note above implies, but there is a disconnect between surnames and DNA. We believe we have tested three family lines that got their "Coffey" name from Edward's line, but did not inherit his y-DNA. And there are also five family lines that got Edward's y-DNA, but not his Coffey name. Let me explain:
Annister Coffey (abt 1708, VA) was a daughter of Edward, and she had an illegitimate son in about 1735 named James. The evidence says this son had a grandson named Lewis M. Coffee that was born in 1798. And we have tested two "Coffey" men who match each other's DNA and who clearly descend from Lewis M. by different lines. But their DNA does not match the Edward group, suggesting the hypothesis that they actually descend from Edward's daughter Annister is probably true.
By the way, we have long had great hopes of solving this 275-year-old paternity mystery, and have so far gone through three hypotheses as to who was James' father - all testable by DNA. And on the third try, we got it right! DNA tests show that James father was someone in the family of one James Samuel (1690-1759), and court records suggest it was James Samuel himself. You can learn more about it if you click here on James COFFEE.
Jane Coffey (1760, VA) was a great-granddaughter of Edward, who passed on her Coffey name to son Jordon, but of course not the y-DNA. And indeed, the y-DNA of the tested descendents do not match the Edward Group. Initial speculation was that the father might be the man she later married, named Fitzgerald. However we now strongly believe the father was a "Taliaferro", based on upgraded DNA tests and on the presence of Taliaferro men in the same county where Jordon was born.
Lucinda Coffey (born about 1830-35, NC) may offer a story similar to Jane. She is a GGG granddaughter of Edward. The parentage of her son Jasper Pink Coffey has long been subject to question. And the DNA test on Jasper's GGG grandson, Ben, indeed came back "no match" to the Coffey DNA. However Ben then decided to upgrade to "67-markers", to see what would turn up - and the result came back with a large number of excellent matches (up to 67 out of 67 markers) to the "Estes" family! Further, census data showed there were MANY Estes families living in Johns River, Caldwell County, NC, as neighbors to Lucinda and her father, Enoch. And Ben now even has a prime suspect, one "Joseph Estes" who lived next door.
And next, William Coffee Berry (1796, NC), and his sister named Mary Coffee Berry, are firmly believed by genealogists to be the children of Rice Coffee (a great-grandson of Edward), via a relationship with Elizabeth (Fields) Berry, who at the time was married to Bradley Berry (or was his widow, accounts vary). This was apparently an open secret. The genealogy is supported by William's middle name "Coffee", by reported recognition in Rice's will, and by an autobiography written by William himself. Anyway, the living descendents surnames are "Berry", but as expected the y-DNA was "Coffey". Case proven.
And then there is an "Adams" connection. This dates back to 1888 in Wilkes County, NC, when Charles Robert Burke/Adams was born. (His mother originally gave him her surname of "Burke", but in later years he changed it to his stepfather's name of "Adams".) The family had been searching for the identity of his birth father, and the DNA clearly offers "Coffey".
And next, there is a Boone Family connection, but the details are still elusive. This involves a line of Boones that descend from the family of Israel Boone (the elder brother of the famous Daniel Boone). The Coffey and Israel Boone families were neighbors in North Carolina, and at some point one Boone line acquired Coffey y-DNA, yielding a match on 64 out of 67 markers.
And finally, we find a Holt Family connection - - or actually two of them. First, we found a person named Holt who tested an excellent match to the Edward group. Then, in the process of exchanging notes with that person, we found ANOTHER researcher had documented a Holt/Coffey family connection, but it proved to NOT be connected to the first tested Holt family. Then we found someone from the second line to test. If you want the LONG version of this story, along with speculation and clues, click here on Holt Connections.
Hugh Group: Hugh Coffey is believed to have entered America in the early 1700's. So far we have three project participants who trace their ancestry to Hugh, or who are shown by the DNA test to be related to Hugh. These participants are clearly NOT genetically related to either the Edward Group or the Peter Group.
Mills Group: This group is descended from Lilburn Coffee, who was born in Tennessee in 1821. Family stories said that Lilburn was actually a "Mills" who was adopted by a Coffee family, and the Y-DNA test proved those stories to be absolutely true! FTDNA reported that Lilburn's descendants had matches in the "Mills" Surname Project, but none with any other Coffey Project members. (Given their "very recent" Coffey origins, we consider the descendents of Lilburn to be the most modern Coffey Clan in the world!)
"Northern USA" Group: Most of the Coffey families that can trace their ancestry back to the 1700's, show that their ancestors entered through Virginia. But there is another group of eight clearly related men who trace their ancestry back to the 1800's, and most entered either through New York, Connecticut or Philadelphia, or are found in northern States. Three of them say their families came from County Cork, so probably they all do?
And this group includes two "non-Coffey Surname" members. One BRANCKY involves a known adoption. In the other, a descendant of one Charles ROSS, born 1864 in MS or LA, tested an excellent match to this group. Charles' descendants have long questioned his name – they now speculate he may have been a Coffey, who had some trouble with the law, and changed his name to "Ross"?
"County Meath" Group: For the first time we have found a DNA matched group (now with ten members), one of whom traces a significant genealogy within an Irish County. Therefore we have the "County Meath" group. They match each other, but do not match any of the other Coffee/y groups. We also perceive that there are two sub-groups here: One “Group A” of 5 are closely related to each other. A second “Group B” of 5 are closely related to each other. But the relationship between the two groups appears to be much more distant.
And there’s another wrinkle. The “Group B” has some nice matches to “Egan/Keegan”, that are actually closer than their matches to “Group A”. Further, the “Egan/Keegan” additionally have two separate more distantly related groups. And they also come from counties near Meath and Westmeath.
And these differences may be explained: This group is flagged by FTDNA as descendants of "Niall of the Nine Hostages", after a 5th century Irish warlord who supposedly left behind thousands of descendants. It suggests that all these Coffey, Egan and Keegan families have common origin and have been in the general area for hundreds of years, and their y-DNA has been gradually evolving in different directions.
Note: Of the 10 tested men, 4 were tested by Ancestry.com. There is not a perfect alignment between the markers tested by FTDNA, and those tested by Ancestry, so there are gaps in the data match. But there is no doubt they are related. Of the 4, we have 3 who have given permission to enter their data into to our project. So there remains one match named “Anonymous”.
"Irish Coffey" Group: Yes, we now actually have two members who still live in Ireland. Fintan lives in Dublin, Sean lives in Galway, and Fintan suspects their common family roots are in County Westmeath. They match each other but, unfortunately for those Americans wishing to establish DNA ties to the Old Country, they don't match any of our presently tested USA members.
"Unknown" Group: There are several persons that the DNA test shows are not related to any other individuals or groups in the Coffey/Coffee database. Maybe future DNA tests will offer matches?
Dr. Russell Coffey: There is one person in our "Unknown" group who is rather special, and new participants may want to sign up just to discover if THEY might be the ones related to Russell?
Russell is interesting, because he lived to the age of 109, and was one of the last three surviving veterans of WWI. His DNA is also interesting, because he is the only "Coffey" in DNA Haplogroup "I" - - almost everybody else named Coffey/Coffee is Haplogroup "R", the most common European group. (We'll talk more about "R" later.)
Group "I" is found throughout Europe, but at low frequency - - it is somewhat more common in Scandinavia. Hey, Scandinavia is where the Vikings came from, and they were rampaging through Ireland more than 1000 years ago! One can imagine where Russell got his "I" DNA?
WE'RE RECRUITING MORE MEMBERS:
We hope that other males with the Coffey or Coffee surname (or other variations) will join our project, including descendants of more recent immigrants to America and perhaps more of those still living in Ireland. If you're not sure about joining, tell me (FredCoffey@aol.com) what you know about your family history. I can then engage you and other "Coffey Cousins" in a discussion about what you (and we) might learn from your test.
VIEWING THE RESULTS:
We have two ways of looking at the results of our testing. To see the basic data, click here on DNA data. (If you find that hard to read, try clicking on Data.pdf to see a more readable and printable PDF version.)
However if you want to see the DNA results in conjunction with the known ancestry of the participants, we offer another approach. You can see it at Participant genealogy and summary (and again, clicking on Genealogy.PDF will get you a PDF version), but please read the following explanation first:
The latter chart shows the male ancestry of each participant, as far back as his genealogical research has taken him. If the participant has serious doubts about some of the connections, these are marked with a "?" on the connecting line. (Remember almost everyone on this chart is obviously named "Coffey" or "Coffee", so only first names are shown. Also, if known, birth year and state are shown in each box.)
We have chosen one individual in each Group as a "Reference" - - usually this is the person who has taken the most detailed DNA test and/or has the most confidence in his family tree based on genealogy research. And the column of numbers to the far left shows how closely the DNA of each participant matches that "Reference" person in his group.
You will likely want to do additional reading on what the test results mean. But for the moment just know that the tests come in five versions - - 12-marker, 25-marker, 37-marker, 67-marker, and 111-marker. And if there is a match on at least 11 markers out of 12, or 23 out of 25, or 33 out of 37, then the participants are likely related. If we include a person in one of our "Groups", then his DNA results match at least on this level.
Just to pick one example, the chart shows that I (L. Frederick Coffey) match the "Edward Group" reference by a score of "25/25" and "66/67". In other words, on the 25-marker test I had a perfect match, and on the 67-marker test I was only out by one marker. And that single marker is known to change more easily than most of the others. These people really are my cousins!
The chart is segmented by Group. Except for Peter and Edward (and their ancient Keogh/Kehoe/Kaho cousins), there is no meaningful genetic match between the groups. You can look at the DNA data and confirm this.
HOW CAN YOU LEARN MORE?
You can find all sorts of general information on DNA testing for the Surname Projects, and how to interpret the results, if you visit www.ftdna.com.
We also have a web site, and you're welcome to visit the Coffey Cousins site, which links to several additional resources. We publish a newsletter, which you can read at http://coffeycousinsclearinghouse.org. Historical newsletters are available at http://www.coffey.ws/familytree/CCCNewslettersOnline/, and this includes an index with over 29,000 entries.
THE CHESLEY COFFEY CONNECTION
I'm going to begin to address some more complicated issues in the remainder of this report. If we've reached the end of your attention span, you're welcome to stop reading:
On the bottom half of the genealogy chart for the Edward group, you will see a group of people who believe they are descended from a grandson of Edward, named Chesley, born in the early 1700's. There is actually a great deal of uncertainty about Chesley, including serious doubt that he even existed. You will see that our chart shows an alternative line that bypasses Chesley completely - just in case!
We haven’t yet found any DNA clues that allow us to sort out the Chesley problem.
THE ELI COFFEY CONNECTION
Similar to the above, note two men on the genealogy chart marked in pink to the left of their names. Meldon and Kevin both descend from Eli Coffey, and they have a perfect 37-marker match to each other. And they are the ONLY Coffey men with an allele value of "17" at Marker 13 – every other Edward descendant has "18". One might think this could be an identifying marker for the "Eli" line, except one other Eli descendant (me!) has the more common "18". One idea was that maybe both Kevin and Meldon were descended from Willis, who was the origin of the mutation to "17". But both Kevin's and Meldon's genealogy look solid, so we're puzzled. Maybe it's just coincidence?
THE "MARKER #27 = 12" CONNECTION:
There are three men marked with a violet band to the left of their names. The tested men (Gordon, George and Danny) trace their ancestry back to three men (Joel, Franklin and Jackson) who were all next-door neighbors in the 1850 census for Russell County, KY.
All three are very solid "Edward Group" matches, EXCEPT all three differ by showing a value of "12" at Marker 27. Everybody else in the Edward Group has an "11" for this marker. The easiest way to explain this is if Joel, Franklin and Jackson are brothers, and the mutation to "12" occurred in their unknown father. And Coffey relationships in that time period do abound with unproven assumptions and unclear clues.
The resulting discussion among Coffey Cousins now runs to pages and pages of theories, clues and opinions – far too much to cover here. If anyone wants to jump in, let me know and I'll put you in touch with our discussion leader.
The Two Gordons: Oh, the astute observer will note that there are TWO "Gordon Lee Coffey born 1941 in KY" entries on adjacent lines. This is NOT an error. They are two different men.
ARE TERRY AND RON CLOSE COUSINS? ALSO DAVID AND KENNETH?
Another interesting comparison is with L. Ronald Coffee versus Max Terry Coffey. They are "Edward" group, but are the only pair within that group that match at Locus 9. Both have gotten the 67-marker test, and they are very solidly part of the Edward group, and no other unique matches are evident. As with the previous case, perhaps this one match is an indication of an unidentified common ancestor between them and Edward? (A dotted line is used on the genealogy chart, to indicate where the researcher is focusing her effort to seek the possible link.)
(A similar "pair" is for David Randall Coffey and Kenneth Cyrus Coffey. They are the only ones with a matching "36" at marker #34. But they already know they have a unique common ancestor in Edward's grandson, Edmund S. Coffey. The mutation from the more common value of "37" most likely occurred in Edmund, and this value of "36" may now be a marker indicating likely descent from Edmund? We'll be watching for confirmation!)
This is another of the more complicated issues, and if we've reached the end of your attention span you're welcome to stop reading.
The issue is that FTDNA reports a genetic distance of "five" for cousin "Raymond Allen". However if you look at the Data table, only three markers have changed. So how do they calculate "five"?
The first half of the answer is easy. If you look at Locus 11 (DYS #392) you see that Ray's result is "11", and the Edward Reference is "13". That's a difference of two, and counts as two changes. (Whether or not it should count as two is arguable.)
The hard part is that we show a genetic change for Locus 12 (DYS #389-2), but the score for both Ray and the Edward Reference is "28". How does FTDNA count this as a change?
The reason is that Locus 10 (DYS #389-1) is considered to be a sub-segment of DYS #389-2. Since DYS #389-1 dropped from "12" to "11", there has to be an offsetting change elsewhere in DYS #389-2 in order for the overall score to remain the same at "28". Got that?
The bottom line is that Ray (versus Luther) has a genetic distance of "four" on 12 markers, and that would be a stretch in claiming a relationship. However on the next 25 markers there is only one more difference, bringing the total to a genetic distance of "five". That is within a reasonable range for claiming a relationship.
COMPLICATIONS: GENETIC DISTANCE
There's another strange result with Kenneth Ray Coffee. FTDNA's basic comparison says his "genetic distance" from Luther is "six", because there are two single changes plus a difference of four at DYS #576. However, after consulting with Bennett Greenspan at FTDNA, we conclude the change at #576 should be treated as a SINGLE change. This makes Kenneth a clear member of the "Edward" Group. (Note the FTDNATiP calculator apparently also recognizes this as a special case, and it shows him as having a high probability of a common ancestor with Luther.)
ABOUT GEORGE AND ERNEST:
If you look at the data for George and Ernest, in the Edward Group, you will see that they each have a genetic distance of "2" from our Reference, and this clearly ties them to the Edward Group. However they both also claim to be descended from "Joel", a possible great-grandson of Edward, and herein lies a possible problem: When George and Ernest are compared with each other, the genetic distance rises to "4" - - and that big a number has to shed some doubt on whether they really do have Joel as a common ancestor. More examination of the paper trail is needed.
MORE ABOUT THE "NORTHERN" GROUP:
The people presently shown in this group are excellent DNA matches. However there is an additional person, Alan Dale, who also has a tenuous match to this group. The FTDNATiP says, based on the 37-marker test, there is a 50/50 chance of a common ancestor with Robert Eugene if you go back 24 generations. Perhaps Alan's family branched off in really ancient times, or there have been an exceptional number of mutations? We've left Alan in the "unknown" group for now, but will continue to look for evidence.
DEEP CLADE TESTS:
This is another of those topics that may go beyond the attention span of some of our readers:
Recently individuals from two of our Coffey groups have elected to get what FTDNA calls a "Deep Clade" test, which is designed to explore ancient origins. This is not going to help us put names on our genealogy – we're talking about "ancient" as being in terms of thousands of years. We're looking back to before the use of surnames like "Coffey" became common.
As humans evolved, there have been mutations in the y-DNA, and people fall into various "Haplogroups" which are indicative of their ancient genetic origin. We already knew, based on the original DNA tests, that almost all of the Coffey groups in our DNA project are in Haplogroup "R1b1a2".
The "R" group first appeared perhaps 30,000 years ago, in central or south Asia. This soon divided to yield an "R1" group, believed to date back about 20,000 years. And a bit later there was an "R1b", which is now found with high frequency in Western Europe. One of the highest R1b frequencies is Ireland - where the frequency of R1b is greater than 90% - although R1b is not believed to have evolved there.
Finally, after further mutations, by about 5000 to 8000 years ago we are up to our "R1b1a2".
The "Deep Clade" tests take this a bit farther. We now know that our "Edward" group is "R1b1a2a1a1b5". And the one we call the "Northern USA" group is just one step different with "R1b1a2a1a1b4".
These results are determined by testing for a "SNP" (Single Nucleotide Polymorphism) in the DNA, i.e., markers that are distinctive for each Haplogroup. As an alternative to the above "R" labels, some prefer to speak simply of the test results. Both of our Coffey groups test positive for the SNP "P312" (i.e., they are "P312+"), which defines the "R1b1a2a1a1b*" haplogroup. The Edward Group is positive for only one subclade under that – they are “L176.2+”.
The Northern Group is different, only in that they are "L21+", i.e. positive for L21, another of the subclades under P312. They are negative for all further subclades under L21.
Details are evolving rapidly, and it is likely that the nomenclature will continue to change, and that estimates of when and where these sub-groups evolved will become available. I haven't seen an estimate of when these two Coffey groups diverged, but you need to be thinking in terms of a few thousands of years, not a few hundred. And if our common male ancestor COULD be found, one can be sure he was not yet using any variation on the name "Coffey".
If any members of our other y-DNA groups are interested, they are invited to get the same test. But note that there is no reason for more than one person from each group to be tested. Given the closeness of conventional DNA match, what applies to one will apply to all in the same Group.