How well do DNA testing services’ analytical tools handle DNA matches for highly admixed people?

It is fair to say the online genealogy community has been set alight with the launch of new research and DNA analytical tools from AncestryDNA and MyHeritage. I can understand and appreciate the excitement. MyHeritage’s new AutoCluster and The Theory of Family Relativity (ToFR) offerings have enabled my to smash through a handful of my most stubborn genealogical brick walls.

It shouldn’t come as a surprise that the new analytical and research tools have generated quite a bit of commentary online: ranging from euphoric to harsh. While the merits of ThruLines, AutoClusters, and ToFR are being discussed and debated, I am going to share my thoughts from a perspective that isn’t being discussed within the wider genetic genealogical community.

How well do DNA testing services’ analytical tools handle DNA matches for highly admixed people?

What do I mean by “highly admixed people? I mean individuals whose genetic inheritance and ancestry is evenly spread across more than one different ethnic group. In my case, that would be a DNA admixture that is evenly spread between African, non-Jewish European, and Jewish European. I’ll leave my smaller amounts of Native American and Asian (Central, Southeastern, and Far Eastern) DNA to one side for now.

For the purposes of this article, I will be referencing AncestryDNA, MyHeritage, and FamilyTree DNA. I haven’t tested with 23andme. If you have tested with 23andme – and are highly admixed – I would welcome hearing your thoughts and/or experiences in the Comments Section.

One caveat: I will not be referencing ethnicity percentages. Ethnicity percentages are not the focus of this article.

I will begin with my experience with AncestryDNA.

One thing has consistently bothered me about AncestryDNA. It has taken the roll out of its ThruLines to bring what has nibbled me about the service into sharp relief. AncestryDNA certainly “sees” my European and Jewish Ancestry in the form of its ethnicity breakdown, and the sheer volume of Jewish and European-descended DNA cousins on my match list. However, it has rarely – and I mean rarely! – facilitated common ancestor matches. This is despite the fact that I have a large and well-researched tree. Ancestry DNA has readily matched me to other African American cousins, either through its DNA Circles and Shared Ancestor Hints. I can’t say the same has happened with my Caucasian DNA cousins – even when we’ve worked out how we are related through paper trails.

In a recent message to AncestryDNA, I stated that I felt (and I stress the words I felt) that there was a virtual One Drop Rule at play, or a kind of DNA apartheid. I explained that what I meant by those terms was this: whatever coding algorithm AncestryDNA was using seemed to focus on my African American DNA to the exclusion of everything else. Please don’t misunderstand the point I am making. My African DNA is important. However, in terms of weighting, it accounts for just over one-third of my Ancestry. Yet, Ancestry’s analytic tools predominantly focuses on just this ancestry.

Below is something that occurs more often than not when I go into AncestryDNA to look at my matches, as well as matches for the many AncestryDNA kits that I manage:

I am greeted by six thumbnails of beaming, happy, Caucasian DNA cousins. So, at this level, AncestryDNA is affirming something I know already…I have quit a bit of European ancestry.

However, when I look at the various DNA stories and Migration Stories that AncestryDNA provides as historical overviews of the external/societal forces that impacted my ancestors, this is what I get:

Notice anything in the three historical overviews? Not yet? They are solely for my African American ancestry. 

I should have, at the very least, a Jewish Diaspora story for my Jewish DNA. I should also have multiple European stories that cover my colonial Quaker, Irish, British Isles, and Ulster Scot ancestral migrations. These ommissions are glaring.

So why is this imporant, and why does this irritate me so much? Well, despite how it happened, I have white ancestors. Numerous white ancestors, to be precise. Understanding the connection between my mulatto ancestors, both free and enslaved, and their white forefathers and foremothers is critical. Why? More often than not I can only find documentation for my African American ancestors through their white family members. I can only piece together my family’s broken African American family lines by connecting them through their shared white ancestors. If AncestryDNA’s coding prevents this from happening, that simply throws hurdles in my ability to piece my enslaved ancestor’s fragmented and slavery-disrupted histories. 

For my ancestors who were free people of color, this impedes my ability to understand their origin stories, as well as obscuring their migration stories into the parts of America that would become Kentucky, Tennesse, North Carolina, and Ohio. This is especially true for my ancestors who were free people of color who assumed white identities when they migrated into the places listed above. This group in particular lived in remote places in the early colonial period. There were no colonial offices for them to register births, marriages, and deaths. There were no formal churches to register the same information. Like a number of my formerly indentured Irish, British Isles, and Ulster Scots  ancestors, illiteracy was more the norm than not: so no scribbling of births, marriages, or deaths in the family bible. In truth, there might be family oral histories regarding their ancestry, but those are patchy at best. There is one tool to make the necessary familial connections: DNA.

To be clear on another point, I am not saying that AncestryDNA is racist. It’s sad and disheartening that I even need to say that, however, considering the times we currently live in, I need to say this just to be clear about it. 

What I am saying is AncestryDNA needs to re-examine the coding that drives its matching algorithm, which, in turn, drives the various analytic tools on its service.

MyHeritage sits at the other end of the spectrum. Its DNA and research tools, thusfar, sees all of my admixtures. Nowhere is this more apparent than its AutoCluster tool. It is for this reason I have been able to obliterate some brick walls that were decades old. MyHeritage’s ability to “see” and understand my heavily mixed admixture is so much better that I will be doing the majority of my DNA analytic work on its service. Whatever coding it uses to drive its matches my reports and DNA analytic tools works with my research, and not against it.

As for FamilyTree DNA? It sits somewhere between AncestryDNA and MyHeritage. Personally, I find FamilyTree’s Chromosome Browser to be a stand-out DNA analysis tool. It is something to use regularly.

If you too are heavily admixed, I’d love to hear your thoughts, both positive and challenging.

And, if you work for AncestryDNA, my standard invitation still stands. I am more than happy to consult with you on this. This is something that needs to be addressed.

Confirming Jason Futrell of Rich Square, Northampton County, NC as my 4x great grandfather

The new suite of genealogical and DNA analytical tools from MyHeritage has literally knocked my ancestry out of the park these past few days. The online service’s The Theory of Family Relativity Tool and AutoClusters obliterated four of my most stubborn ancestral brick walls…and confirmed the identity of a man I have long suspected to be my 4x great grandfather.
This article is about that all important ancestral confirmation.

I began researching the life of Bug Frutrell of Rich Square, Northampton County, North Carolina, my enslaved mulatto 3x great grandfather, six years ago. As per my usual practice, I began working on a family tree for the white enslaving Futrell family. I rapidly fell into the genealogical quagmire that is the Futrells of Rich Square. A handful of family lines were well documented and researched. Most lines were not. The one lineage book that was digitized and available online was riddled with errors. Like other parts of my ancestry, the Futrells used the same five or six names for their sons and their daughters, resulting in individuals who were either conflated into one person, or confused for one another.

I reached a point where I was on the verge of giving up when a white Futrell DNA cousin, Becky, reached out to me on Ancestry to introduce herself, and to say hello. Like me, she was curious to know who our common Futrell ancestor was. Understanding that my connection to the family came through the enslaved Bug Futrell didn’t phase her in the least. She took that in her stride. 

Becky shared a copy of a family lineage book that is no longer in print: The Futrell Family. Simply put: this book was awesome. It remains the most accurate book about the Futrells that I have accessed. The paper trail backs up its findings. In short order, I had a workable and usable Futrell family tree.

Armed with an accurate tree, I had a relatively short list of enslaving Futrell men in Northampton County to research who would have been of a suitable biological age to have fathered Bug. This kind of research involved finding Wills, estate inventories, slave insurance policies, deeds that cited enslaved people, contracts involving hiring out enslaved people – any relevant document that would contain the name Bug between 1819/20 and 1865, when Bug would have been emancipated via the Emancipation Proclamation.

I found Bug cited in an appraisal for Jason Futrell of Northampton County after a month or so of searching. Jason was the right age, and lived in the right place, to be Bug’s father.

DNA triangulation and segmentation showed I was more closely related to Jason’s white descendants than any other group of white Futrell descendants. I shared more cMs and longer chromosomal segments with Jason’s descendants than any other Futrell descendants with roots in Northampton County, North Carolina. 

So…I added Jason as Bug’s father on my Ancestry tree – and waited for shared ancestor hints to pop up on Ancextry.com. I was waiting to see shared ancestor matches with Jason’s other children, as well as the descendants of his maternal and paternal aunts, uncles, and grandparents. I waited. And waited. And waited. There was nothing. Which was strange, considering the sheer number of white Futrell descendants I matched on Ancestry with deep roots in Northampton County.

Fast forward six years later.

I hadn’t logged into MyHeritage for almost two years. My research focus has mostly been on Ancestry. However, I was hearing so many excited reports about MyHeriage’s Theory of Family Relativity and AutoClusters that I had to check them out for myself. 

Who was the subject of the first theory? You guessed it – Bug Futrell. Who was his predicted father? Jason Futrell. 

Looking at my AutoCluster Report, I quickly saw my Futrell cluster. In and among my 15 Futrell DNA matches were six white Futrell cousins I knew from Ancestry. I went on to check a staggering number of messages in my MyHeritage inbox. There, nestled among my messages, were system reports from five years ago pretty much telling me that Jason was Bug’s father.

I was estatic. I then became furious. Why had MyHeritage been able to report something five years ago that AncestryDNA had never even suggested?

I wrote a pretty scathing correspondence to Ancestry outlining what I’ve shared above. While I haven’t received a reply, my Futrell connections on Ancestry finally appeared when I logged into Ancestry the next day. I haven’t added all of them. The four images below will suffice:

I hope part of why I’m still salty about this is relatively easy to understand. Confirming Jason as Bug’s father enables me to understand how, and through whom, I match descendants of the other mulatto Futrells in Northampton County. It strengthens my ability to find members of Bug’s immediate and extended family. It is critical in understanding Bug’s family’s story. Denying me that confirmation denied me the ability to even try to reconstruct the family history of the enslaved mulatto members of the Futrell family.

This week has highlighted a few things for me. It is my feeling – and I stress the words my feeling – that Ancestry had a significant amount of work to do in two areas:

  1. The way its DNA matching algorithm handles comparing DNA for people whose ancestry has off the chain endogamy; and
  2. The way its matching algorithm handles comparing DNA for people who are heavily admixed.

Me being me…I have both. MyHeritage seems to be better equipped to handle both. The brick walls that crumbled this week involve both centuries worth of endogamy, as well as a high degree of relations between Europeans, Africans, and Native Americans occurring over centuries.

On the upside, I can now begin the work of connecting the dots between the different mulatto Futrell family groups in order to understand how Bug connects to them.

How to Compare Unrelated People on AncestryDNA

This article from the DNA Genealogy blog absolutely rocks!!

I’ve tried this exercise out, taking it for a thorough spin…and it really does work.

The good peeps over at DNA Genealogy walk you through the process step by step. Please surf on over to http://www.geneticgenealogist.net/2016/04/how-to-compare-unrelated-people-on.html?m=1 to learn how to do it.

new2bancestor2bdiscoveries1

Sample AncestryDNA landing page screen grab

 

How to get YDNA haplogroup from AncestryDNA results

MorleyDNA.com Y-SNP Terminal Subclade Predictor image

MorleyDNA.com Y-SNP Terminal Subclade Predictor. This image shows a snippet of my results. It turns out the I have a newly designated Haplogroup – the origins of which are hotly contested by scientists and academics.

This post from the DNA Genealogy website steps you through the process of filtering the YDNA information buried deep within AncestryDNA’s autosomal DNA results.

It works!  I tried it for myself. And there it was – the same haplogroup two other DNA testing services had already provided.

I virus-checked the free software that needs to be downloaded and found it to be safe. It’s always worth remembering that we should always virus check software – especially free software.

There’s a few steps to go through. However, I recommend giving it a go.

I hope there will be a follow up for extracting mtDNA data from AncestryDNA autosomal DNA tests.

To extract your own YDNA results from an AncestryDNA austosomal DNA test, read the following article and follow the easy steps:

How to get YDNA haplogroup from AncestryDNA results

http://www.geneticgenealogist.net/2016/01/how-to-get-ydna-haplogroup-from.html?m=1

The Roanes of Virginia: 2 families with the same surname. Are they related or not?

What could possible be confusing about two immigrant families coming from the same region in Europe and landing in the US around the same time?  When it comes to pre-Revolutionary War Era Roane family…there’s plenty.

One group of early 18th Century Roanes were Scots-Irish in their origins, descendants of the northern Irish landowner of Scottish origins, Archibald Gilbert Roane.   The other Roane family hailed from England, descendants of Charles “The Immigrant” Roane.

Untangling A Right Genealogical Mess

As I’ve previously written, these two men were not directly related to one another. If I had the power to correct every single Roane family tree that shows Charles as being the father of Archibald, I would do it in a heartbeat :o)

Many years ago, like any newbie amateur genealogist, I figured countless online family trees had to be correct. I mean, they had been published for years – long before I began my own genealogy adventure. What wasn’t there to trust? The majority of these tress had merged both of these Roane family groups into one family. I took the information they contained as gospel. About a year later, I realized just how wrong these trees were.

It’s the only time I have ever had to delete an entire family from my tree and start again from scratch. However, it taught me a valuable lesson: never, ever take what’s in family trees as the gospel. The fact that most of the trees didn’t have citations or documentation should have been a clue. You live and you learn.

Part of the confusion, admittedly, was the realization that the American authors of these trees didn’t understand the distinction between England, Scotland and Ireland. They didn’t know the history of the UK and Ireland either.  A basic knowledge of a country or region’s history can guide your genealogical research. Historical knowledge can raise red flags. That’s what happened to me with the Roanes back in their respective countries of origin. The authors of these incorrect trees assumed that two people born around the same time with the same name were one in the same person, regardless of where they lived. So a Robert Roane, who clearly lived and died in Midlothian region  of Scotland, with a wealth of christening, marriage, death and property records to show that he was mainly domiciled in Midlothian, was presented as being one and the same as a Robert Roane who lived and died in Sussex County, England – and was sometimes a resident of London. The Scottish Roane was a wealthy Scottish merchant who was part of the sphere of influence for Queen Mary of Scotland. The other, a wealthy courtier and advisor to Queen Elizabeth I.

A myriad of official records and personal accounts clearly show these were two different men. Yet, there they are, on one online tree after another, where they are shown as being one in the same person. An extra layer of complexity.

Now there may or may not be a shared common ancestor between these two very different Roane families. Both claim descent from the ancient Norman/French noble house of de Rohan. And, of course, there is no proof of this. Until their direct male descendants do DNA testing, I can only take this as speculation that has yet to be proven.  For the time being, I’m treating them as two different families.

Two Families with The Same Name In Colonial Virginia

It gets even more confusing for these two families in the American colonies. The Scots-Irish Roane settled in two places when they first arrived in the American colonies: Pennsylvania and Virginia. The English Roanes went straight to Virginia. Virginia is where things gets interesting. Both Roane families were in the premier league of Virginia society. And, like any Division 1 family, they married into other Division 1 families. Which means these two Roane family groups became connected through marriages – marriages with families like Ball, Brockenbrough, Henry and Upshur/Upshaw. This, in turn, means that autosomal DNA from families like Ball, Henry, Brockenbrough and Upshur/Upshaw runs through both the English and the Scots-Irish Roane lines in the US. It makes it a challenge to know which Roane family group you’re a descendant of if you don’t know your direct line of Roane ancestors.

It’s especially difficult and confusing for African American descendants of either of these families (which I’ll get to directly).

Both Roane family groups were large slave-owning families. Generations of Roane men, from both family groups, fathered children with slaves. Pinpointing which male from either of these Roane families fathered your African American Roane family line is like looking for a needle in a haystack. And the shared DNA thing is just an added curveball, plain and simple.

There are a LOT of African American descendants from both these houses who are actively wanting to know which Roane family group they belong to in order to pinpoint their direct line and then uncover the identity of the man who fathered their line. I receive numerous emails from African American Roanes every week asking for advice, insight and my input into helping them along this path of discovery. Trying to untangle the knot of oral family history (which isn’t always correct), complicated family inter-connections, and interpreting DNA match results takes patience.

Working Through My Roane DNA Matches

I only made my breakthrough in identifying William Henry Harrison Roane as the progenitor of my Roane line through an exceedingly lucky break. Quite literally, through the luck of the Irish!

There were three Scots-Irish Roane brothers who arrived in the American colonies. Two headed for Virginia while one, the Reverend Andrew Roane, lived in Pennsylvania.  Luckier still, Andrew’s descendants largely remained in Pennsylvania. Which means this was as pure a Scots-Irish Roane line as anyone is likely to research in the US.  I say pure because this line never married into the same families as the Virginia Scots-Irish Roane lines before the outbreak of the American Civil War. In terms of DNA testing comparisons, this one Roane line is gold dust.

My DNA matches with Andrew’s descendants proved there was a genetic link between me and the Scots-Irish Roanes. After months of meticulous and laborious DNA results triangulation, I could eliminate a direct line of descent from the English Roanes (although I shared DNA with them through those pesky marriages to other prominent Virginia families). I could also whittle down my direct line of descent within the two Virginia-based Scots-Irish Roane lines until I finally hit my direct line.

I’ll give you an example. In terms of DNA matches, all of William Henry Harrison Roane’s living descendants who DNA tested with AncestryDNA were my closet matches (1st cousins ‘x’ times removed. They typically show as 4th cousins on DNA testing services).  William’s siblings’ descendants were the next closest series of DNA matches (2nd cousins ‘x’ times removed. They typically show as 5th cousins on DNA testing services). The descendants of William’s uncles and aunts were one more generation removed (3rd cousins ‘x’ times removed. They typically show as 6th cousins on DNA testing services).

My DNA matches with English Roanes has been consistently more removed than any I share with Scots-Irish Roanes. My English Roane DNA matches go from 8th cousins to ‘distant’. This suggest that we’re not linked by English descended Roane men. We’re linked by society ladies from the same family who married into both Roane families. And these unions had to have happened two to three generations prior to the birth of my 4x great-grandfather, William Henry Harrison Roane.  Which my Roane family tree actually shows. This makes sense. Families like Ball, Brockenbrough, Henry and Upshur/Upshaw had arrived in the colonies generations before either of the Roane family groups arrived.

A Little Bit About Triangulating DNA Matches

A note about triangulation. You have to compare your DNA matches with male descended lines and female descended lines. Which means you have to have a fully worked up family tree with both male and female lines to gather the surnames you’ll need to search on. For instance, it was a 50/50 shot whether it was William or his father who fathered my direct Roane line. It was comparing my DNA matches with his mother’s Henry family that clinched it. In order for me to have Roane and Henry DNA, I had to be descended from a child of a Roane-Henry union. That would be William – whose descendants were my closest DNA matches compared to any other line of Roane descendants.

Knowing the lineages of the women in your tree is every bit as important as the male lineages. And nowhere is this more important than DNA match triangulation.

Andrew Roane is my Roane family litmus test.

Some Roane Family DNA Matching Interpretation Tips & Tricks

So, if you’re an American Roane descendant reading this (and especially an African American Roane descendant), here are some suggestions:

  • See if you match with a living descendant of Andrew Roane with DNA results posted on the various DNA/family history sites.
  • If you do match a descendant of Andrew Roane, and that match is between the 3rd to 6th cousin level, then you are more than likely a descendant of the Scots-Irish Roanes.
  • If you don’t have a match with Andrew Roane’s descendants, then you are more than likely a descendant of the English Roane family group.
  • I don’t have any 7th cousin level matches, so can’t offer any interpretation for that result.
  • Once you’ve determined which Roane group you belong to, keep comparing your matches until you find a line that matches you more recently in time than any other. The chances are high that this is your direct line. There is a caveat:
    • Work back quite a few generations to ensure there are no linking families in your family line (two sisters marrying two brothers, or a pair of cousins from one family marrying siblings or cousins in another family). It happens more than you think. And this will skew your DNA cousin matching results. Think of these as potential false positives.
    • You will need to keep searching until you find family lines that aren’t connected by a linking family. This one is important. When I researched the ancestral lines of two people who married an aunt of uncle of William Roane, I discovered they were descendants of my Harling ancestors. Which made these two people my genetic cousins. Put simply, I had a set of Harling cousins marrying a set of Roane great aunts and uncles. This meant I had to completely ignore their Roane descendants in terms of making DNA comparisons. There simply is no way of saying ‘just look at the shared Roane DNA and ignore the Harling DNA’. I wish there were.
  • Trust me, the temptation is just to great to force your DNA match results to fit the information in your tree. Only refer back to your family tree for surnames to search for in your match results. However, forget about the specific individuals in your tree.
  • Don’t let your family history assumptions influence your match interpretations. Actually, forget everything your family oral histories have passed down. You have to triangulate like Dr Spock from Star Trek: dispassionately, focusing solely on the results.

Another consideration for African American Roanes is where your ancestors lived in Virginia. If you know where they were born, lived and died-  that, in and of itself, might give you some clues as to which Roane group you belong to if you’re getting Roane DNA matches.

The slave owning Scots-Irish Roanes are most strongly associated with King William and Henrico Counties in Virginia.

The slave-owning English Roanes are most strongly associated with King & Queen and Gloucester Counties.

Essex County & Richmond are tricky – both Roane family groups had land holdings and slaves both of these places. I’m still working on an overview of which family group was where in Essex County. For instance, which of the two families were in Indian Neck, Tappahannock, etc. This will also be a clue.

Ultimately, it will be a combination of finely-researched genealogy, DNA testing and patient and thorough triangulation that can unlock the mystery of which Roane family group you belong to.

With this is by no means definitive, also look at popular names in your family, especially names that keep recurring for family members born between 1800 and the 1880s. I’ve discovered that this was one way the formerly enslaved people of color indicated which white family line fathered them. Male names like Charles and Robert were popular for black and white descendants of the English Roane line. Henry was a popular first and middle name for black males who were descendants of William Henry Harrison Roane. Patrick, Spencer, Anthony and Wyatt were also popular male names associated with Scots-Irish Roanes.

This Research Has Inspired Something Even Bigger

The complications with Roane family genealogy fits quite nicely into one of my main goals in my genealogy adventure. And that is building one of the biggest online, public, slavery-era family trees for African Americans. One that is fully researched and trusted. It’s a project that I’m currently applying for grant funding to realize. I’m thankful that my current family tree is already a resource for African Americans researching families associated with Virginia and the Carolinas. However, I’d like to go bigger, deeper and further back in history. And that requires full-time research commitment for quite some time.

In the meantime, I continue to chip away. And always look forward to sharing whatever I find along the way.

So I’m starting to explore AncestryDNA’s ‘Shared Matches’ tool…

I’ve recently started to explore AncestryDNA’s  ‘Shared Matches’ feature. By recent, I mean only 48 hours ago. Explorations like this makes my inner geek happy. I’ve tinkered with digital stuff and technology for years, so tech like this isn’t daunting to me. I fearlessly dive in to see how things work…or least try to figure out how things like this tool work.

I offer this caveat up-front: I’ll be covering the ‘Shared Matches’ feature as it appears on the new Ancestry / AncestryDNA site. So please don’t be thrown or confused when looking at the screen grabs. They will look very different from the old version of the site, if that’s the one you’re still using.

I’m just exploring and working things out at the moment. There is an end game for this exploration. I hope I can make some in-roads into my Irish and Ashkenazi Jewish genealogy. Yes, that’s right, I have 3 of the most challenging ethnicities to research when it comes to genealogy: African-American, Irish and Central/Eastern European Jewish. Three ethnicities that have undergone a worldwide diaspora with some of the most challenging records to find. That’s me. If there is a genealogy higher power, s/he must be laughing.

What better way to finely hone my pending approach to tackling my unknown Irish and Jewish ancestors than a thorough understanding of how this feature works with my known African-American cousins and known European cousins? I can then apply this insight to DNA cousins I have yet to find common ancestors for. If I can understand the strengths and weaknesses of the ‘Shared Matches’ algorithm and results, I can have a more informed  approach to understanding the many Irish and Jewish DNA cousins I have on AncestryDNA. I have a staggering number of each.

That’s my working premise at the moment.

Ok, with that out of the way, let me show you how I’ve been exploring this AncestryDNA tool.

Below is the standard AncestryDNA family matches landing page. No surprises here for those of you familiar with the service. And yes, your eyes aren’t deceiving you. I really do have 49 pages of DNA cousin matches.

image showing my AncestryDNA family matches landing page

My AncestryDNA family DNA match landing page. Please note that I have will be protecting the identities of my DNA matches in this post. Click for a larger image

Now I have a LOT of cousins with connections to what was the Old Ninety-Six County of South Carolina (this county was dissolved to create the following counties: Abbeville, McCormick, Edgefield, Saluda, Greenwood, Laurens, and Union counties, parts of  Spartanburg County; much of Cherokee and Newberry counties; and small parts of Aiken and Greenville Counties). I know exactly how I’m related to a number of these 3rd, 4th and 5th cousins. There’s a group of us who are very, very active genealogy researchers and share information of various family groups on Facebook.

I connect to these cousins in a myriad of ways. We are the descendants of enslaved Africans, free people of colour and Quakers who left England and Scotland for Antrim and Ulster in northern Ireland, made their way to Pennsylvania and later the former Old Ninety-Six. Knowing how we’re related tells me a bit about how ‘Shared Matches’ works.

And, in a way, my Edgefield family heritage is a good one when it come to understand this DNA matching analysis tool. Whether Quaker, formerly Quaker or African-American, my Edgefield ancestors and relations married within the extended family for generations. So many of my Edgefield-connected cousins and I, regardless of ethnicity, are related to one another a few times over (so far, the winner is a cousin I’m related to at least 4 different ways). Why is this important to note? There are few single sets of common ancestors when it comes to the Edgefield side of my family tree. Which makes pinpointing isolated common ancestors a bit tricky. I’m going to find the same pattern for my rural, agricultural Irish ancestors (who married within a clan structure) and my Ashkenazi Jewish ancestors. Edgefield, it turns out, will be a great genetic genealogy proving ground. It really is as complicated and intricate as general genealogy, much less genetic genealogy, gets.

So, I want to see how many DNA cousins there are who also share a connection to Edgefield. So I typed this search term in the search box as shown in the image below:

image showing how to filter for a specific place on AncestryDNA

Image showing how to filter for a specific place on AncestryDNA. Click for a larger image.

And this is what I got:

image showing how to filter for a specific place on AncestryDNA

Click for larger image.

Turns out I have 78 at the moment. I’m sure if I have added the other counties created from Old Ninety-Six, this number would have been much, much higher.  For this exercise, I want to solely concentrate on Edgefield.

Roughly a third of these 78 matches are cousins that I share a common set of 18th Century Quaker x-times-many great grandparents who lived in Pennsylvania. These ancestors didn’t live in Edgefield themselves – but had descendants or extended family members who did. The only reason they appear in the search results above is down to the fact that, like me, they are researching the whole family and not just their own direct line. I already had their ancestors in my tree and they had mine in theirs.

So I’ve placed this group of cousins to one side. For this exercise, I’m solely focussing on cousins whom I know I share a common set of Edgefield-born ancestors.  It’s always a good idea to grapple with a small sample size when dealing with something new; and something as complicated, complex and intricate as genetic inheritance. It’s like learning how a car engine works for the first time. You wouldn’t tackle how the engine works as a whole. You start with how two parts of it relate to one another and work together, and then add a third engine component and a fourth and a fifth until you finally understand how the whole engine works. This is how I’m approaching the Shared Matches tool.

78 people is just far too many to begin to unravel the mystery of this DNA analysis tool.

So I started hunting around within these results for someone who would match me and only a handful of other DNA cousins on the service within these Edgefield results.

The hunt is on for a DNA match who matched me as well as other Edgefield-based DNA matches. Click for larger image

Trial and error: the hunt is on for a DNA match who matched me as well as other Edgefield-based DNA matches. Click for larger image

So I searched around until I found the DNA cousin above. I’ll call her ‘Mary’. 

When I clicked on the ‘Shared Matches’ link on her page, this is what appeared:

Click for larger image

Click for larger image

Mary matches with me as well as 3 other people. Now this is a sample size I can work with when it comes to analyzing the Shared Match tool!

To re-cap, the 5 of us (Mary, myself and the 3 people who share DNA with Mary and I), share ancestors who lived in Edgefield. Which is exactly what I wanted. Now I have to unravel how, exactly, we’re related. Why are there only five of us -and not all of the other people who match me for Edgefield? Why do my known Edgefield cousins (from families like Matthews/Mathis, Holloway, Settles, Williams, Dorn, Ouzts, Peterson, Timmerman, Harlan/Harling, Gilchrist, Borum, etc) match me, but not Mary or the 3 matches she and I have in common. What family line do the five of us share that none of my other Edgefield relations do? Understanding this not only fills in a genealogy information gap. It will give me a sound insight into how the Shared Matches tool works. Only when I understand how this tool really works can I begin to extrapolate and apply what I’ve learned to other groups of shared matches.

I have a feeling that this Shared Match tool is AncestryDNA’s compromise offering for not having a Chromosome Analysis tool like the ones available from Family Tree DNA and Gedmatch.

image A chromosome analsyis i ran on DNA couin match group on Family Tree DNA

A chromosome analysis I ran on DNA cousin match group on Family Tree DNA. Click for larger image.

AncestryDNA’s position on not having a chromosome analysis tool is entrenched. Like many others, I think it’s a bad call.  Knowing the DNA segment lengths you share with DNA matches can provide critical insights. I have nicknames for parts of my chromosomes that I match others on on Gedmatch and FTDNA: names like my Roane segment or my St. Clair segment, my Josey segment or my Matthews/Mathis segment. Or I name them by region: those are my Arab chromosomes or my Central Asian Chromosomes or my Jewish Chromosomes. On FTDNA and Gedmatch, I don’t even have to know the name of person to see how we’re related. More often than not all I need to see is which chromosome segment we match on and share.

While Ancestry’s Segment Match is better than nothing, and will ultimately yield results, it’s not really a substitute for a chromosome anlaytics tool.

With that said, a few things have already piqued my interest with this group of DNA Shared Matches. Mary and two of the other matches have a 100% European ancestry as reported by AncestryDNA. One person has an Ancestry that’s as mixed as mine. Which, initially, tells me that the shared Ancestor pair we have in common is most likely European. And looking at the major ethnicities of Mary and the two other European-descended matches, this common ancestral pair has the highest likelihood of roots in England, Wales, Scotland or Ireland.

And I’m very excited about the guy with an ancestry as mixed as my own, who I will call Joe. Why? Because I know Joe. I know how we’re related on our known white and black Edgefield lines. The common ancestors Joe and I share can’t be shared by Mary and the other 3 people in this shared match result. Which means I can exclude all of the ancestors and relations that Joe and I share in common when it comes to identifying the common ancestral pair that links us to Mary and the 3 others in these results. Somewhere out there is a new white family name I have yet to find. One that Joe and I don’t share with all of our other known Edgefield cousins who have taken the AncestryDNA test to-date.

This is the benefit of working with a small match pool. It narrows the parameters which results in a narrower field of inquiry. And if this is all beginning to sound like a forensic, CSI-esque kind of experience?  Well, it kind of is. Again, it makes my inner-geek happy.

The next step is to dive into the family trees for these matches…if their trees are public. Heck, if they even have them at all. in this case, only one of the trees is private. So I have 3 trees to work with. Which is kind of a lucky break. Trust me, 3 trees to work with was more than I could have hoped for.

The next step will be applying what I will be learning about this tool to other Edgefield family match groups that are larger. And when I have a finely-tuned understanding of this tool? I’ll start applying it to the Irish and Jewish DNA cousins where absolutely nothing is known in terms of ancestors we share in common. That is my end-game.

 

AncestryDNA: So what does it take to get a DNA Circle?

So I’ve previously shared my frustrations with the whole Ancestry.com DNA Circles thing. Namely, the fact that I have a distinct lack of what AncestryDNA refers to as DNA Circles.

For those of you not in the know, DNA Circles on Ancestry.Coms DNA testing service purportedly go beyond finding a common ancestor with your DNA matches. These circles are meant to link you to additional AncestryDNA members with the same common ancestor…thus creating a Circle of people who are all related. Nice and simple, isn’t it? :O)

Given the size of my tree and known DNA matches for my family lines such as Sheffey, Roane, Harling and Josey –  I shared my frustration about the fact that I didn’t have a single DNA Circle on Anctery.com.  I felt (and still do) that this was a legitimate gripe…and a gripe shared by many using the service, especially those with African American lineages.

Two months ago two names suddenly appeared on my AncestyDNA landing page. Now, the sting in the tail was these two names appeared as “New Ancestry Discoveries” and not as DNA Circles. And, of course, neither name was familiar to me. Then, just as suddenly as these two names appeared, they disappeared just as quickly.

So you can imagine my surprise when these two individuals appeared once more today.

ancestrydna01

click for larger image

I have no Medders or Altmans on my family tree. So, in order to determine how these two people could conceivably relate to me, I had to do some digging.  And this is what I discovered:

I clicked on the link for John Smith Medders.

ancestrydna02

click for larger image

I then clicked on “See Your Connection” in the right column…for obvious reasons. And got this:

click for larger image

click for larger image

This left me none the wiser about who John Medders was or how we might be related. So I clicked on the “Relationship” link, hoping this might shed some light.

What this gave me was a list of Ancestry.com members I shared varying degrees of DNA with:

click for larger image

click for larger image

Well, one thing became quickly apparent: I was definitely in the realm of the Medder family. Each and every individual was a member of various Medders family groups on Ancestry.

The second thing that quickly became apparent was that I had a solid DNA match with two individuals – the same two individuals that are shown in the third image in this post.

In order to “see what I could see’, I selected the “View Relationship” for both individuals. And that’s when things quickly clicked. I’m only going to show one of the relationships to illustrate the discovery.

click for larger image

click for larger image

The surnames of Flowers, Gregory and Moore are exceedingly popular surnames in America. However, taken collectively, and with roots in Pennsylvania, and then the Carolinas, I knew exactly what family in my own tree these names related to: the Harlan / Harling family. Yep, another Quaker family connection via the Quaker Harlan family.  The Harlan / Harling family had married Flowers, Gregory and Moore for nearly three centuries: first in England and then northern Ireland. And continuing such marriages in Pennsylvania and then in the Carolinas.

In this instance, Hannah Flowers b. 1722 (a cousin many times removed), married a Joseph Ashton. Their daughter, Hannah Ashton, married William Thomas.  Hannah and William’s son, William Jr, married Celia Alice Gregory (yet another Quaker cousin through the Harlans). The Meddars family shown for my two DNA connections above are descendants of William Thomas, Jr and Celia Alice Gregory.

So, at the very least, I am a distant cousin to at least John Smith Medders.  I may yet be a cousin of Mary Ann Altman. At the moment, I haven’t come across any familiar family names in the family trees I’ve seen for her.

So, while these two DNA matches don’t have a single Harlan or Harling in their tree (yet!), I get the connection.

I don’t get the lack of DNA Circles though.  Of which I still don’t have a single one. Go figure.

click for larger image

click for larger image