On Saturday, June 11, 2016 at 3:59:16 PM UTC-7, Matthew Langley wrote: > On Saturday, June 11, 2016 at 8:53:34 AM UTC-7, Vance Mead wrote: > > I have some questions regarding the use of DNA in genealogy, in view of the recent postings on the subject. Much of the recent discussion, it seems to me, has been based on straw men and false dichotomies. "DNA can show your ancestry back to Adan and Eve or else it is completely worthless." No doubt this is due, at least in part, to the unrealistic claims of companies trying to sell a product. > > > > As far as I can tell - and I am certainly no expert - DNA testing can supplement documentary evidence in a few ways: > > > > - It was used to prove that the skeleton found in Leicester was in fact that of Richard III. > > > > - It could show if Granny had a secret she never told Grandpa about who was Dad's father. > > > > - It could show if people with the same surname are really related. In my case, there were several Mead families in Oxon/Bucks/Herts in the 15th and 16th centuries. One question: how far back is this reliable? For example, there are proven descendants of Thomas Mede of Henley on Thames, Oxon, born about 1550, and of Richard Mede of Watford, Herts, born about 1515. They are probably unrelated, but could DNA show a theoretical connection between them some 15 to 20 generations ago? > > > > - I have seen how it has been used to show that two people are fifth cousins, or whatever. Is this only in the male line, or in all possible relationships? > > > > I can see how DNA could be used to support documentary evidence, much as archeology is used to supplement historical evidence. Could someone give a brief summary of how it is used in genealogy? For someone who is more at home in the map and large document reading room than in a forensics laboratory. > > > DNA is absolutely useful for genealogy. Ignore the thread you saw, many genealogists are using DNA effectively. > > There are three types of DNA tests: > > Y DNA: > > Representing the paternal line only Y DNA is *immensely* useful for genealogy. For example as a Langley I've been Y DNA tested. I dead end on my ancestor with a Langley that lived in the late 1700s in South Carolina. I've Y DNA matched other Langleys that trace to people in that region, one who sold land to my ancestor. Via DNA I now know without a doubt those Langleys share a paternal ancestor to mine. > > Based on the amount of markers tested and how close the match is I can get a rough idea how closely we relate too, such as within 200, 300, 400, 500 years etc. > > When you map Y DNA to paper trails it's amazingly useful and can confirm or deny believed relations. > > Further if a Langley in England Y DNA matches me in the future I can confirm we share paternal ancestry and that will help me find my immigrant ancestor and where they came from in England. > > > mtDNA: > > like Y DNA but purely maternal lines, this is far less useful since mtDNA mutates far less often than Y DNA. This means that you may match someone else on your mtDNA but that doesn't narrow it down to a genealogical time frame depending on the level of testing. You can usually get an idea whether two people share a maternal ancestor within 500-1000 years or so. This made the Richard III testing valuable since his mtDNA matched a maternal descendant of his believed maternal line, basically the odds of that being a coincidence are pretty insane, this is how they confirmed he was indeed Richard III. > > > aDNA: > > Autosomal dna is immensely useful. You pretty much have to ignore any segments that match below 7 cM (maybe even 10 cM). But 10 cM+ is something like over 99% chance of being a relation within a genealogical time frame. > > The true value comes in finding a cluster of people who *all* match the same segment of aDNA. This means all those people got that same segment from the same source somewhere. So say you have 3+ people who share a segment and you find that all have a shared ancestor in their tree, it becomes very likely that specific segment came from that ancestor for each person (the more people in the cluster obviously decreases the odd everyone shares another set of ancestors that is unknown). This means you can identify segments as from certain ancestry, then match other people on that segment with no known tree overlap, you can find ancestry further up that line via this method. It takes a lot of work but has been invaluable for many genealogists, including myself. To give an example of using aDNA. My mother's paternal ancestry is Bryant. I've found multiple people who share the same Bryant ancestry in the 1800s. Via enough matches 4-6 I've found some segments I've identified as Bryant DNA, at least Bryant to the point where we all share ancestry. Not too long ago I found a cluster of people, roughly 4 or 5 that all match each other fairly closely and they all share Bryant segments. They also matched my kits (mother, uncle, and cousin) at pretty high levels, 4th cousin range-ish with some fairly large segment matches. Turns out they all descend from a woman who was living next door to an ancestor of mine in Georgia, John P Bryant, in 1850. Her husband died (or disappeared) yet she still kept having children, the same ones that DNA match my family. As it turns out my ancestor John P Bryant was already cheating on his wife with another woman, and by 1860 abandoned his first family (within a year or possibly even before my ancestor was born, his youngest of children with his first wife) and is living in Tennessee with his second family. Further those children from the widow stop being born within the year John ran off to Tennessee. Combine this with the fact that John P Bryant's sons are probably too young to be knocking up the neighboring widow and it becomes very likely those children are illegitimate children of John P Bryant and the high level of matching suggesting that generation amongst multiple matches across multiple kits (3 on my side roughly 4 or 5 on that branches side) A mystery that might have been impossible to solve without DNA.

<snip> > mtDNA: > > like Y DNA but purely maternal lines, this is far less useful since mtDNA mutates far less often than Y DNA. This means that you may match someone else on your mtDNA but that doesn't narrow it down to a genealogical time frame depending on the level of testing. You can usually get an idea whether two people share a maternal ancestor within 500-1000 years or so. This made the Richard III testing valuable since his mtDNA matched a maternal descendant of his believed maternal line, basically the odds of that being a coincidence are pretty insane, this is how they confirmed he was indeed Richard III. <snip> For a practical application of mtDNA to proving a theory (actually two theories) see... Perry Streeter and Mike Morrissey, “Mitochondrial DNA (mtDNA) Testing of Probable Eighth Cousins Confirms Their Kinship and Proves Their Matrilineal Lines of Descent from Aeltje Cornelis (Cool) (Van Couwenhoven) Stoothoff; Simultaneously Proving Two Long-Held Theories Developed Sequentially Through Conventional Research; Mary (Updike) Richey of Greenwood, Steuben County, New York Was a Heretofore Unidentified Daughter of Roliph-5 Updike (John-4, Lawrence-3, Johannes-2, Louris-1 Jansen); Eleanor5 "Nelly" Lane (Gizebert-4, Mathias-3, Gysbrecht-2 Laenen, Matthys-1 Jansen) Was the Heretofore Unidentified Wife of Roliph5 Updike,” New Netherland Connections (April, May, June 2005; Volume 10, Number 2). ...which is available online as... http://freepages.genealogy.rootsweb.ancestry.com/~streeter/NNC/mtdna.pdf I am working on similar project right now!