Monday, April 2, 2012

"Finding Your Roots with Henry Louis Gates Jr." - DNA in The Third Episode

Last night PBS aired the third episode of Finding Your Roots with Henry Louis Gates Jr. which featured journalist Barbara Walters and educator Geoffrey Canada. The episode revolved around investigating the guests' paternal roots by tackling the challenges of discovering the original surnames of their fathers' ancestral lines. Geoffrey's was obscured mostly due to his extremely limited knowledge of his father's family. Although Barbara had previously had her family tree researched, her immigrant ancestor's name change had not been discovered.

After discovering that Geoffrey Canada's paternal line was descended from Thomas Cannaday, a slave owned by a Charles Cannaday of Franklin County, Virginia, Geoffrey's Y-Chromosome DNA was analyzed to determine if his great-great grandfather Thomas may have been fathered by his slave owner. Not surprisingly, Canada does possess a European Y-Chromosome Haplogroup. In fact, he apparently shares his haplogroup with Kevin Bacon, Robert Downey, Jr. and Harry Connick Jr.  Since the majority of men in the Bacon DNA Project, the Downey DNA Project and the sole man in the Conrick DNA Project (possibly a variant of Connick) possess the R1b1a2 Y-Haplogroup (the most common in Europe), it is likely that Geoffrey's Y-DNA Haplogroup is also R1b1a2. This piece of information is not enough to solve the puzzle though, since there is no way to know for sure when that European ancestor entered Geoffrey's direct paternal line, so Dr. Gates' team identified two great grandchildren of Charles Cannaday and asked them to take an autosomal DNA test to determine if their great grandfather is the same man as Geoffrey's 3rd great grandfather. If the working theory was correct, that would mean that these Cannaday descendants are Geoffrey's 2nd cousins twice removed and would be expected to share about .781% of their DNA (the same amount as third cousins). In this case, there is a chance that the autosomal test could be ineffective due to the fact that only about 90% of third cousins have enough shared DNA to be detected by this kind of test. Although it would certainly be worth a shot, this scenario could be stretching the capabilities of autosomal DNA testing just a bit. Even if they were to have a small amount of shared DNA, without extensive testing of other descendants, it might be difficult to say with certainty that the DNA came from this potential shared ancestor rather than another, especially since we know both families' ancestors lived in the same geographical area. In my opinion, a Y-STR DNA test on Charles Cannaday's direct paternal descendants (or his father's, uncle's or brothers' direct paternal descendants) would be more conclusive because it would determine if Geoffrey is carrying a Cannaday Y-Chromosome or not. Unfortunately, the theory did not get to be tested anyway because, in the end, both of the Cannaday descendants identified by Gates' team declined to participate in DNA testing. Dr. Gates, with deadlines looming, was probably forced to leave the question unresolved for now. Many of us genetic genealogists know how it feels to track down the perfect candidate to test out a genealogical theory, only to find that the person identified is unwilling to take a DNA test. It is always disappointing, but given the luxury of time, perhaps his team wouldn't have stopped there. Since the local Cannaday researcher said that she had 3200 Cannadays in her database and that it was still a very common name in the area, the team could have either looked for descendants of Charles Cannaday's father or asked his great great grandchildren to test. (I have found the younger generations are often more willing.)  In either case, the potential relationship to Geoffrey would have been one more step removed, thus decreasing the chances of a match even further, but still worth investigating.

Obsessive genetic genealogist that I am, I just couldn't let the Y-STR idea go, so I searched for a Canada/Cannaday DNA Project to try to determine if there were any Cannadays already tested, against whom Geoffrey's Y-STRs could be compared. I was initially surprised to find that no project of this name exists, however I did find a Kennedy DNA Project that lists variants Canaday and Canady. Upon further investigation, I found that there are 15 Canadas, 12 Canadys, six Cannadys, seven Canadays and one Cannaday who have tested at Family Tree DNA (they may not all have taken a Y-STR test). If you look at the Kennedy DNA Project's Results Page, you will find some of them there. There are seven participants of interest - four named Cannady, one Canady, one Canaday and one Canada. Six of these seven fall into the haplogroup R1b1a2. (The one who does not is identified as being descended from a Charles Canaday, but there is no information supplied as to birth or location.) It may be that none of these are directly descended from the direct paternal line of Charles Cannaday of Franklin County, Virginia, but just for kicks, it would be interesting to see if Geoffrey's Y-STRs match any of these participants. I wonder if Dr. Gates' team investigated this avenue.

Like last week, geneticist Joanna Mountain from 23andMe appeared on the episode - this time to explain the Y-DNA analysis. Interestingly, in this segment there appears to be two spots where there are inaccuracies in regard to this DNA test. The first one occurs starting at 41:15 when Dr. Gates asked Dr. Mountain, "So a Y-DNA analysis, in other words, is what people in the barber shop talk about the paternity test?" and she (appears to have) answered, "It sure is."  The exchange seems awkward and I believe this mistake was made in editing. I am sure that Dr. Mountain would never imply that a Y-DNA test is the same thing as a paternity test since it unquestionably is not. (I think what happened is that her comment was not actually in response to that questions, but was pasted in from another part of the segment.) A traditional paternity test is based on autosomal STRs markers, like CODIS . While it's true that a Y-DNA test could informally be used as such, in reality it cannot tell us if two people are father and son, only that they come from the same direct paternal line. (They could actually be brothers, uncle/nephew, first cousins, second cousins, etc...) The second mistake was probably merely a slip of the tongue on the part of Dr. Gates, but still worth clarifying since genetic genealogy is so new to many of the viewers. Starting at 43:35, Dr Gates explained what the results of this Y-DNA test mean for Geoffrey, "We do know without a shadow of a doubt that a white man fathered your female slave ancestor [bold mine] and entered your family line." I'm sure he meant to say "impregnated your female slave ancestor" since it was actually Geoffrey's male ancestor that was fathered by a white man. We know this because the Y-DNA only follows the direct paternal line. By definition, women cannot be introduced into this line because they do not inherit a Y-Chromosome. I don't mean to call anyone out on these mistakes, but I do think it is important to explain and clarify in an effort not to confuse any viewers new to the concept of genetic genealogy.

The next time DNA was mentioned in the episode was when Dr. Gates asked Barbara what percentage of her DNA she thought was of Jewish origin and she guessed 99.9%. An admixture test, which appeared to be FTDNA's Population Finder (included with the Family Finder test) revealed that she is actually ~91% Middle Eastern (Jewish) and ~9% European (non-Jewish). Dr. Gates noted that this exercise with Barbara illustrated that DNA analysis is revealing that "our family trees are more diverse than we had assumed."

This concept is proving true for African Americans as well. "In the African American community, genetic evidence of our rainbow-colored roots is challenging long-held assumptions about what it means to be black," Dr. Gates explained.  To illustrate this, reminiscent of last week's scene in the barber shop, Dr. Gates asked several students from Geoffrey's school to estimate how much of their DNA is African, European and Native American. After swabbing their cheeks, the students were surprised to find from their admixture tests (which appeared to be from 23andMe) that their ancestry was more diverse than they first imagined.  After learning that a portion of her DNA was European one student said, "I didn't know that I was European.. now I don't consider myself just black." Gates emphasized that these tests "deconstruct the notion of race" and reveal that "we are all mixed up".

I'm very pleased that DNA was again a vital component of this episode, but a little disappointed that they weren't able to use it more conclusively to determine Geoffrey's paternal ancestry. I guess that is realistic though and better than implying that genetic genealogy is always easy. By the way, can you believe Barbara Walters is 82?! She must have some pretty darn good DNA! I thought it was kind of funny that her family's original surname Waremwasser sounded an awful lot like Gilda Radner's Baba Wawa. I look forward to next week's episode with Kevin Bacon and Kyra Sedgwick. I hope they perform autosomal DNA tests on the couple to see if Kyra's fear is true and they are indeed "kissing cousins". See you then!


  1. Hi CeCe,

    Nice summary of Episode 3 ... and more questions!

    In regards to your use of Y-Haplogroup terminology I would appreciate some clarification and additional elaboration.

    You said: "... it is likely that Geoffrey's Y-DNA Haplogroup is also R1b1a2"

    As many know, R1b1a2 (SNP R-M269) is not a Terminal Haplogroup. Whereas one such as R-M222 presently is.

    Saying R1b1a2 (I prefer ... R-M269)is like telling a Genetic Genealogist Geographer a descendant being researched is from California where they want more precise information such as this: R-M222-92672-9920 and the son of R-M222-61603-2236 ... it's easier to connect the dots like that and visualize patterns and family groupings and linkages ... it's like a Genetic Genealogy HTML code!

    No, they did not reveal the Terminal Haplogroup for G. Canada. If the Head Genealogist for Episode 3 knew it ... then either she, Canada or Gates likely made that decision. Perhaps the program last night will get some people volunteers to test to assist Mr. Canada. Stopping at 2 only NOs left me scratching my wondering.

    If they had Canada's Y-STR DNA profile at FTDNA they could have run it through one of the commonly used Haplogroup Predictor programs, do a FTDNA $29 SNP test or a broader Deep Clade test. Did they? ... that would be interesting to know if they have his Terminal Haplogroup ID.

    What are the known Terminal Haplogroups for Harry Connick Jr., Kevin Bacon, Robert Downey Jr.?

    With the Autosomal DNA test at 23andme I do not believe they have the capabilities of IDing all the current Genealogy SNPs under R-M269. I know they can do larger SNPs such as R-M222 but not all the Terminal SNPs in the R-L21 Phylogenetic Tree.

    Since you are an advisor to 23andme ... perhaps you can clarify this and report back to us on the exact Terminal R-SNPs 23andme can currently ID and compate that to FTDNA and

    I also share your concerns about the limits of Autosomal comparison testing of 4th and more distant cousins. I think getting a clearer answer directly from the Big 3 ... 23andme, FTDNA, is important here.

    If you really want to get a person's DNA ... as we all now you can try a family member (ala the CIA and the son of Bin Laden) And thre is a variety of ways to get DNA these days as the CIA well knows trained mosquito's, garbage harvesting, forensic touch analysis, water bottles, postage stamps, licked envelopes, etc etc etc.


    BTW Cannaday and Kennedy are pretty close in my Scots-Irish Soundex algorithmic code!

    1. Hi George,
      We can only work with the information that the producers gave us, which is what I have tried to do here. We do not know anything about Downey's, Connick's or Bacon's haplogroups (yet). That is why I attempted to determine the most likely one from the Y-DNA Projects of their surnames. Since I did not see any overlapping haplogroups across those projects and the Kennedy Project other than R1b1a2, I am guessing that that is probably the right one. It is merely an educated guess and I could be proven wrong if further information is revealed in the future. There is absolutely no way to know what their terminal SNPs are, so I didn't even try to determine that.
      As far as the autosomal analysis, I had the opportunity to go back and review the episode as it is now online, and discovered that the Cannaday descendants identified were actually two generations closer to Charles, the slave owner, than Goeffrey. As a result, I adjusted my wording above to reflect 2nd cousins twice removed, which is a bit better for autosomal testing than 4th cousins, although still not foolproof.
      I'm sure many of us would be happy to assist Mr. Canada with a STR analysis of his Y-DNA, but I doubt he needs our help. Gates' team may have already gone down this road with no conclusive outcome, so chose to leave it out entirely. I was merely commenting that this would be a good avenue of exploration for those who may not realize what goes on behind the scenes and/or be familiar with the different types of DNA tests/analysis available.
      Thanks for your comment.

  2. Hi CeCe,

    I believe a top class Genetic Genealogist has to be highly skeptical, very inquisitive, constantly scratching their head, be a great librarian and know where almost everything is in the Internet realm and the bricks and mortar realm, willing to learn and try new methods and .... maybe not always think to brightly as the brightest crayon in the Mensa coloring box.

    So I graciously take exception when one says a Monday Morning Genetic Genealogist (or a Monday Morning potential client) or one of your readers has to limit themselves and "only" work with the information provided to them on a Genealogy TV program such as "Finding Your Roots" or "Who Do You Think You Are?"

    Such is the case with Mr. G. Canada from New York and others just beginning or muddling around with their Roots

    Canada ... what a cool person and surname to do research for. As noted on the program, even he was not familiar with his surname variants as he easily could have been found at:


    For free at this site, he could easily found a statistically significant geographic cluster of persons in Virginia to research. Maybe some are in his line and maybe some are not. At the early point ... consider all. That site alone would have been a big step for him!

    Then he go to the free part of Intelius and see there were 98 Cannaday in Virginia clustered largely in Franklin Co. It does give some indications which lines to search and which courthouses to search and use the services of a local low cost $10/hour or so researcher.

    A quick Google search would have brought him to the 1800s Charles Cannaday line in VA

    (4)Research on 1850 - 1860 Schedules 1 (Free) and 2 (Slaves) correlating to 1870 Census info and other info would be a goldmine for him as an African American. Did you know that the 1860 Supt of Census almost got Congress to let him enumerate names of slaves in the 1860 Census Schedule 2 ... but Congress turned him down? There are foolproof methods I have developed to overcome that shortcoming and it could be done on a massice scale by and others if they desired to do so.

    He could have then went to the Franklin Co. Interactive GIS website and ID about 3 to 5 Cannaday men to "first" talk about "family genealogy" before even mentioning "family genetics" or his race. He could have also discovered that Frankin Co. is the moonshine capital of the world and people there are most likely highly skeptical of newcomers and strangers. So, that's probably not one of the better places in the county to ask someone to help you out and give up there DNA. Maybe a E or W coast place like Berkeley, CA or Cambridge, MA would have been a bit better to wrangle someone to take a DNA Test.

    He could go to and got a free tel# versus paying AT&T $1.99 for a 411 call.

    Since he is researching his Male / Paternal line I would advise he go to FTDNA. Order up 2 Y-DNA 37 Marker STR tests ($149/tesr) and have both sent to his home. Canada would take one test and mail the other with a Prepaid envelope back to FTDNA to the person in Frankin Co VA is was working on getting tested. If need a $25 or $50 Visa Gift Card always helps!

    I appreciate your clarification on the cousin relationships in regards to Autosomal Tests. We all need a cousin chart tattoo on our arm! Mine is on my left arm.

    I still contend that 23andme is not up to speed with analysis of recently discovered (past 24 months) Terminal Y-SNPs such as FTDNA is.

    I think as a 23andme Ambassador you could dig deeper there and compare the 2 companies on that basis. I see you as an Advocate ... as well as an Ambassador.


    1. I think so too, George, which is why I analyze the use of DNA in this show (and others) so closely, offer my opinion on superior DNA analysis methods and point out inaccuracies. When I said that we can only work with the info given, I was referring to his and other celebrities' terminal Y-SNPs that you asked me about and are unknown to us.
      Of course, Mr. Canada himself could go further with this research if he so desired, using any of the avenues you propose, the ones I propose and/or many others. That is not what I was referring to. As I've said before, we do not know what their thousands of hours of research turned up since we only are privy to a small fraction of that info via the finished show. The genealogy team is a very fine group of individuals I am sure and undoubtedly utilized many, many resources in their research.
      No, 23andMe is not as up-to-date on their Y-SNPs as FTDNA and/or the ISOGG tree. No one disputes that. Y-DNA testing is not their focus and the majority of customers would prefer they use their resources elsewhere. (I know this because we surveyed them and I hear from them privately every day.) I'm not saying that it isn't important, but fortunately we have FTDNA who specializes in this area and does a great job. You're right, I am an advocate for all genetic genealogy and the companies who give us good products.
      Thanks for your comments,

  3. I have such mixed feelings about television shows which cite DNA as evidence! Yes, they let folks know how useful it can be. And none of them seem willing to attempt anything approaching accuracy in exactly *how* they can be useful.

    1. Occassia,
      I know what you mean, but if they can bring them to us, we can educate them. :-) That is why I am writing these posts after each show to attempt to clarify the "whats" and "hows" behind the conclusions reached (and not reached). Without the exposure that these types of shows offer, the databases will never hit that critical mass that we all need and dream of for our research goals to be fully met.
      Thanks for your comment,

  4. Excellent synopsis of the show CeCe and helpful clarifications over the several verbal missteps that occurred. I really like the format of this show and you are correct - even with the errors of fact, it will spark interest in genetic genealogy.


  5. I also enjoy your coverage of these shows, CeCe. Although I am a rank amateur, I am pleased when I notice 'blips' in the shows that you point out - that's another learning tool for me! My hope is that just mentioning DNA testing and explaining its possible value will cause a few of those 'private' matches to change their status and the ones I've emailed to actually respond. And the next person who decides to test actually might be a cousin!


    1. Thanks Kathleen! I have noticed a consistently high "hit rate" on these posts, so I do think it is inspiring interest in DNA Genealogy. Thanks again for reading and for commenting!

  6. Why did PBS fail to publish Barbara Walter's DNA test that established she and her family were 99% Semites?

    1. Hi Ro,
      I'm not sure exactly what you referring to with your question. The Population Finder test from FTDNA estimated that she is 91% Ashkenazi Jewish. Are you aware of another test that she took?
      At the time of this episode, 23andMe did not estimate the percentage of Ashkenazi Jewish ancestry.
      Thank you,