What!? AncestryDNA isn't going to provide segment analysis?! Say it ain't so!!

I have to stop my posts about recent adventures with the NSDAR and say a thing or two about this  recap of the 2014 (and first) Institute for Genetic Genealogy conference in Washington, D. C. by The Legal Genealogist which you can find here brought to us by Judy G. Russell. This is longish excerpt and I would rather use shorter (and less copyright offensive) clips but this is real important, so Judy I hope that you don't mind. Here's what she said in the opening paragraphs of her post about the conference and AncestryDNA's recent decision.

The Institute for Genetic Genealogy — brainchild of Tim Janzen and CeCe Moore — opened Friday with registration and three overview sessions on the testing companies. Attendees got a chance to take a look at information from AncestryDNA, 23andMe and Family Tree DNA in general, with some good general background information being offered.

There weren’t any surprises in that general info — except perhaps the depth of the genetic genealogy community’s unhappiness with AncestryDNA and its decisions (a) not to provide segment data to its customers and (b) to discontinue YDNA and mitochondrial DNA testing and to discontinue even providing links to results of those tests taken at Ancestry. Let’s just say that the unhappiness was abundantly clear during AncestryDNA’s presentation.

OK, do you get that? AncestryDNA will NOT be providing segment data. I thought they were going to. This is a major problem and I really do not understand why they wouldn't want to. We need that segment data.

So it works like this. If you did autosomal testing then you see the list of "cousin" matches. On AncestryDNA you might get to see the other person's tree if they have one there and if you're really lucky you might be able to get a hint about where you and the other person match and who your common ancestor might possibly be. There are problems with this. It is only matching both of your trees.

Here's what I mean. The other person got on your match list because the DNA matched to some degree stated as a percent and a guess about degree of cousin you might be to each other. The tree match is a separate thing and simply a search function on trees and has nothing whatsoever to do with shared DNA. Nothing.

If you want to know which portion of your DNA you share or have in common, you need to look at chromosomes. And that's the "real deal" when it comes to this DNA for genealogy thing. If you can't look at and compare chromosomes then you're just taking someone else's work that your shared match is at that ancestor Ancestry found on both trees. There's no proof. So if one of you has a big error on the tree, and that can happen, you might get an entirely false match.

Chromosome matching is the very best tool when it comes to working with DNA for genealogy. And don't we deserve to work with the best tools available? Sure we do. AncestryDNA, get your act together, man!


The URL for this post is:

GEDmatch is coming back real strong!

GEDmatch.com is my all out favorite tool when getting down to the comparison of cousins DNA results to see where we match. Was recently in touch with one of Mom's strong matches on AncestryDNA and even though we both have good healthy sized trees to look at, neither of us - nor the Ancestry matching function - could see where we matched! 3rd to 4th cousin, and no match on our trees...? Huh? Obvious conclusion: a big fat mistake on one of our trees.

Cousin Michael was the match. He's great to work with, replies quickly and with as much info as he has. When I told him about GEDmatch he popped on over and had his raw file uploaded in a flash. And now that GEDmatch is going again (and looks to be better than ever) his data was finished with the first phase of processing, called tokenizing, by days end and that means that our kits were available for one-to-one comparison. The full processing which allows one-to-many analysis will take longer.

Because that one-to-one comparison was quickly available I could see that his kit and Mom's kit had really juicy matches on three chromosomes. Here's what that looked like.

Chr	Start Location	End Location	Centimorgans (cM)	SNPs
1	99733258	152971219	27.9	6363
6	1227564	9652130	21.3	2620
21	18580589	27602185	16.0	2435

Largest segment = 27.9 cM
Total of segments > 7 cM = 65.1 cM
Estimated number of generations to MRCA = 3.9

Then I took that information on over to my Farrell DNA Project spreadsheet and could see right away that he doesn't fit in there because there were no shared chromosome segments. My Farrell Project peeps do not match anywhere on chromosome 1 or 6, at all! And only one person even has even an inkling of anything happening on 21 and it's not on this particular segment.

So where does he fit? If not the Farrell line, which includes the surnames of House, Biggerstaff and Hartley too, then where on Mom's tree might he match. I went on over to Ancestry to check out Mom's Big Tree so that I could more easily visualize where the match could be and here's what I saw, with the candidate lines in the red box.

Click to enlarge.

The Williams, Price, and Whetstone lines are what we're after. Because Michael matches at about the 4th generation back, a list of the surnames and locations we're looking for is helpful.

Williams, Edwards, Price and another Williams are the surnames associated with Mom's paternal grandfather's line and they're all from Wales. None of Michael's ancestors are from Wales but that doesn't mean that there might not be a connection back before he can see, based on what's not on his tree. 

Mom's paternal grandmother's lines are from Wales - Thomas - but Mom's paternal grandfather's line - Price and Hill- are mostly form Virginia. Michael has a lot of folks from Virginia on his tree so this is a commonality that bears closer scrutiny.

Now we come to the Whetstones. They hail originally from Germany and then probably Switzerland and on to Pennsylvania before the Revolutionary War. I don't see anything hopeful there.

So it looks like we need to take a closer look at the Price and Hill lines from the earliest days of the Virginia Colony forward. And we especially need to look at any lines that aren't filled out to about the 5 generation mark.

Well that's the problem in a nutshell and how we'll proceed. Don't know how much luck we'll have but at least it's a way forward. What I need is some Williams DNA and I know where to find it. Cousin Andrew tested on 23andMe.com and so now I'm off to connect up again and ask if he'd be willing to upload to GEDmatch. It would be very helpful to have Whetstone DNA as well... but I have to figure out how to get that.

Hey, I just stumbled into another GEDmatch tool that helps me get an idea of Dad's DNA! Dad passed before all of this DNA for genealogy stuff got going. But now I see how to get it! I'll blog about that next time when I've had more time to explore this new-to-me GEDmatch tool.




The URL for this post is: http://nutsfromthefamilytree.blogspot.com/2014/08/gedmatch-is-coming-back-real-strong.html

The Thomas and Judah Farrell DNA Project: First try at chromosome mapping and GEDmatch is down

I love GEDmatch and my most serious DNA analysis work is done there. Sure, it's fun to contact DNA cousins at 23andMe or AncestryDNA but I really need a good chromosome browser and GEDmatch is the ticket. But today it's down and right at a critical junction when I badly need the use of its tools and it's going to be down for a couple of weeks. Grrr.

Now that I've moaned and groaned a moment let me tell you what I'm up to. Cousin Rich and I have about a dozen or so people in various states of DNA match. Mom and Uncle Sonny are the most solid match and share the most DNA. They both descend from the two oldest daughters of Thomas and Judah Farrell. (You can read the overview of the Farrells and this project here.) Mom tested with 23andMe and Uncle Sonny and his sister Aunt Mary tested with AncestryDNA. Mom has now tested at AncestryDNA too. Rich and I have kept a list of DNA cousins and some have good trees and some don't. One guy is adopted so we don't know how he fits in. But they all have some matching segments shared with Mom and Uncle Sonny.

Mom and Uncle Sonny are our "benchmarks" because they share the most DNA. Here's a look at the chart GEDmatch whipped up for us showing exactly how they share their DNA. Nice, huh?

Chromosome browser from GEDmatch.

As each new DNA cousin popped up an idea kept forming and reforming in my noggin. I could see the relationships and the stream of chromosomes, but a pattern wanted to form but just couldn't. Ever have that happen? Now let me tell you how I stumbled into exactly what I needed.

First, I saw that there were two terms and that I might have confused: chromosome browser and chromosome mapper. What you see above is a chromosome browser. It is a tool that lets you compare the DNA of two or more individuals to see exactly which segments of DNA are shared. You can then introduce a third person into this mix and that's called triangulation. If you have three people sharing exact (or very close) chromosome segments, that's called triangulation and with solid trees for all three and only one shared most recent ancestor, it's a sure bet that the chunk of shared chromosomes came from the shared ancestor.

A chromosome mapper is a spread sheet that shows you which segments on which chromosomes are shared between many people. In this way you can determine which chromosome segments came from which ancestor with even greater confidence. With any sort of luck, you can then know that anyone with those exact segments descends from that same ancestor. Yes, that's right. If an adoptee who knows nothing about his ancestors also shares those specific segments they descend from that common ancestor.

Chromosome mapping spreadsheet using a Kitty Cooper template with our DNA matches.

Thank you, Kitty Cooper!!!

 

 

You need to know that I have never used Excel before, and it kinda scared me. Looked too complicated. Never mind, because I wanted to do this mapping thing so bad I just jumped in the Excel water. I went on ahead and downloaded a template from Kitty Cooper which you can find here. It's the CSV one. I googled how to use Excel and played with it for a while and found out that I could enter any value in a column and then use Sort and Filter on the Home bar to get the numbers in numerical order. Sounds way, way harder than it is. (Cousin Rich is a whiz at spreadsheets and he'll have a good laugh at my expense reading this! He's in Scotland with family playing golf so maybe he'll miss this post.)

 

Next step was to decide what the column heading should be. You can see the ones I chose but now that I've played with it, think I'll add "number of generations to MRCA" or most recent common ancestor. See post before this for an overview of our requirements and that should tell you why this is important. There are two columns that you don't see here and those are GEDmatch kit number and email. And obviously you can see the attempted redaction of surnames. Not pretty but it works. 

 

There are two names that don't have any MRCA or most recent common ancestor info and they are Stephen and David. One is adopted so he came to our DNA matching party with no tree. What we can now tell him is that he shares some of the DNA that came down to us through the Thomas and Judah Farrell pairing. It could be DNA that came with them from Ireland and thereby came from ancestors "upstream". Or it could have come to him from one of the other of Thomas and Judah's children. We just don't know. But he's real family to us.

 

The second name without a MCRA has a nice tree and ancestors who lived about 10 to 15 miles from where Thomas and Judah lived in what's now West Virginia. We have yet to find our shared ancestor and connect our trees. He's one of us but we don't know exactly how.

 

There's more to this spreadsheet than you can see here. I have some names and GEDmatch kit numbers but not too much more than that ... because GEDmatch is what I need right now and it's down. Again. GEDmatch, I love ya but you're about to drive me crazy!

 

The GEDmatch web site says that they are moving to new servers and once that's done the world will be bright and new because the new servers should increase reliability. It's going to take a couple of weeks. Ugh.

 

The old railroad overpass on the Potomac River near what was once Magnolia, Vest Virginia, home of the Farrell, House, Hartley, and Biggerstaff families who are the subjects of this DNA study.

 

 

The URL for this post is: http://nutsfromthefamilytree.blogspot.com/2014/06/the-thomas-and-judah-farrell-dna.html



The Thomas and Judah Farrell DNA Project: Chasing Chromosomes

This time last year I couldn't have told you very much about chromosomes and now I'm swimming in them. These days I'm focusing my efforts on how various descendants of Thomas and Judah Farrell have inherited certain chromosome segments along ancestral lines. (You can read about the larger project here.)

As we each pursue our various genealogy projects, it often causes us to think deeper about the topic at hand and this sure has been true for me while chasing chromosomes. As regards this project I've stumbled into some conclusions and want to share them. These conclusions are very much based on the way our project is shaping up. Our Farrell project is unique to us but there might be something of interest to others. Maybe.

We know that chromosomes are key to unlocking the science of inheritance. If there's a chromosome match that follows some basic parameters then descent is assured (no matter what the records say or especially what they don't say). But for our project, what are those parameters? Here's our thinking and the parameters we came up with. Just remember that I'm no expert on genetic genealogy:)

1. We need to see the match plus a tree in order to know who the common ancestor is. If there's a match but no tree (or the tree is poorly developed) we've found that we're mostly out of luck, unless we want to invest a lot of time helping the matching person build out their tree. Right now we have only six people who have matched and have identified a common ancestor, but we've just started.

2. Importance of the number of centemorgans, or cMs, in the match: more = better. So we know that cMs is how the match is measured and that more cMs equals a better and stronger match. The general rule of thumb is that under 5 to 7 cMs is probably not worth spending too much time on: glance at it and move on quickly. Over 5 - 7 is called IBD or Identical By Descent to indicate that the matching chromosomes are shared by descent, with some level of confidence. Less than 5 is called IBS or identical by state, meaning that the chromosomes are shared most probably by chance or if by a common ancestor, then that ancestor is way back.

As a side note here, at each generation we inherit only about 50% of our DNA from one parent. Which DNA segments are handed down is close to random. By about 5 to 9 generations you start to lose ancestors entirely leaving no trackable DNA.

3. Endogamous populations are a little different and Magnolia, Virginia, now West Virginia, might be one. It was a small relatively isolated community and there were many intermarriages. An endogamous population is a group or social set in which there's mating amongst a small population and therefore the gene pool is shared and re-shared repeatedly. In this endogamous population situation the rule is to use 10 cMs as a cut off point for IBD and establishing a true genetic connection. This 10 cM cut off could be a good tool in evaluating our Farrell matches.

4. Number of generations to a shared common ancestor. We've seen that when it comes to number of generations to that common ancestor, fewer is better and gives us a better chance at spotting a common ancestor. At the point of 4th cousin it is commonly thought that there's maybe a 50-50% chance of the match being reliable. With each generation the confidence drops some more. We want 4th cousins or closer. (Who doesn't?)

5. In looking at the individual chromosomes where there is a match, the key element is the start and stop locations. If two people have a matching segment of over 10 cMs on the same chromosome with greatly similar start and stop locations, that's a good match for us to pursue.

And what are our "best case" requirements?

* A tree that shows the common ancestor
* 7 cMs minimum and 10 cMs or more is best
* 4th cousin or closer, 5 generations or fewer to common ancestor
* Matches with similar start and stop locations on chromosomes

So there we are. We've found that most matches don't work out because we're thinking that they aren't "good enough." We'll either have to wait for more people to be tested that meet our requirements or invest more of our own time helping matches that follow the four generation and greater than 10 cMs rules build out their trees in an effort to find that most recent common ancestor. We will have to evaluate each match and, using these parameters, decide how much time we can invest to make the match gain more confidence and eventually yield that ever-elusive and critically important Most Recent Common Ancestor.

This is at times difficult work, but so very rewarding.


Resources I like:

Beginner's Guide to Genetic Genealogy
Lots of resources and love the way learning is broken down into manageable lessons. You can pop in and learn on the fly or brush up on a topic. Love it!

Blaine T. Bettinger's Blog: The Genetic Genealogist
It's all good here. Check out the Product Reviews to sort out vendors.

CeCe Moore's Blog: Your Genetic Genealogist
I really don't know how CeCe gets it all done. Her schedule makes me swoon with exhaustion. Use the search box down a little ways on the right to connect with any topic you can think of.

ISOGG Blog: the International Society of Genetic Genealogy
The Source. Check it out.

DNAeXplained – Genetic Genealogy
Roberta Estes's web site is a good read. Subscribe or pop in to browse or search.



The URL for this post is:
http://nutsfromthefamilytree.blogspot.com/2014/05/the-thomas-and-judah-farrell-dna_30.html

The Thomas and Judah Farrell Project: DNA connections using AncestryDNA, Part 2

I need to emphasize again that I do really like AncestryDNA and it's key feature: using the Ancestry Member Trees to find matches and make cousin connections.  The interface is super user friendly and it's easy to navigate when seeking and finding DNA matches. I like that. Of course, I wish I'd find even more matches... don't we all?)

In the last post I wrote about our recent Biggerstaff side project and my longing for a chromosome browser built in to AncestryDNA. Now I want to turn the spotlight on a search mystery that's got Cousin Rich and me scratching our heads. The mystery has to do with the AncestryDNA search function especially as it relates to surname matching. You need to know right here that I have a very limited grasp of the mechanics behind how search apps work and that extends to the AncestryDNA search app as well. I just know what happens when I do this or that.

If you remember from the last post, Mom matches Cousin Joseph because they are both descended from Isaac Biggerstaff (1798-18440). Uncle Sonny is also a descendant of this Biggerstaff line, but not from Isaac Biggerstaff. But Uncle Sonny doesn't match Cousin Joseph. Why?

The answer could be as obvious as Mom and Cousin Joseph sharing Bigerstaff DNA that came right from Isaac Biggerstaff. The DNA shared between Mom and Uncle Sonny could actually be Farrell or House DNA. It remains to be seen and more will be known once Cousin Joseph uploads his raw file to GEDmatch so we can play around with the chromosome matcher utility.

When I found Cousin Joseph's match for Mom I emailed Cousin Rich and he went to see if Uncle Sonny or Aunt Mary also matched him. Joseph wasn't to be found amongst the regular list of matching people so Rich did a surname search on Biggerstaff. Still no Cousin Joseph.

Rich has a good sense of these things and it was Rich who first questioned if the AncestryDNA search function might have something off-kilter going on.

To double check I searched on Whetstone. Used that surname because I'd recently been in touch with a high confidence match who shared our Whetstone ancestors. When I did the surname search, whatta ya know, she didn't show up!

I have no idea why this should be. Is AncestryDNA looking at the same main match list and just searching for surname matches?

Now do you see why I really, really want AncestryDNA to tweek their search function and hopefully making it as good as the search function on the geanealogy side of the house?

 

 

Click here to get to this next search box, below, and enter your desired surname.

 

Late breaking update: tried the Whetstone surname search just now and it worked! But why didn't it before? Now I'm more confused than ever.


The URL for this post is:  http://nutsfromthefamilytree.blogspot.com/2014/05/the-thomas-and-judah-farrell-project_15.html

The Thomas and Judah Farrell Project: DNA connections using AncestryDNA, Part I

Let me say right off the bat that I do enjoy and appreciate AncestryDNA. But people, for Pete's sake, get yourselves a chromosome browser! And maybe take a second look at your search function especially as it relates to surname matching. Otherwise you're fine... but I really need these two things, ASAP. Now here's why I need this, and I'm going to talk about the chromosome browser thing first and in another post address the issues we're having when we use the search function for surnames.

Let's chat about the usefulness of a chromosome browser. What's a chromosome browser, you say? In a nutshell, it's the ability to run a search or matching app to see exactly where in their DNA two people share the same chromosomes. When we're lucky, some segments of those chromosomes  come down through the generations from the shared common ancestor relatively in tact in each person's DNA. A chromosome browser will find the matching segments for you.

How important is this? I've read a genealogy blog recently (wish I could remember which one, dang it) that compared the chromosome browser to the idea of "original source" in standard genealogy. I can use a chromosome browser to find out which chromosomes exactly match Mom and Uncle Sonny's DNA.

Here's a look at the GEDmatch comparison of Mom and Uncle Sonny's matching chromosomes. Mom first tested with 23andMe and Uncle Sonny and Aunt Mary tested with AncestryDNA.  In order to see exactly how and on which chromosomes they match we needed the help of a 3rd party service like GEDmatch. Here are the results for Mom and Uncle Sonny.

 

In order to get this report both parties must upload their raw data file to GEDmatch, and that's no trouble really, but if the other party is reluctant for no particular reason, then you're out of luck. If you have a chromosome browser built-in to the DNA service you are using, you can just go on ahead and see where you match with the other person's DNA, without the fuss.

Here's a link to Ce Ce Moore's blog where she talks about all this and Ancestry's plans to add their own chromosome browser. I can't wait... but there's no release date at of yet. And I really have a personal problem with waiting because I WANT IT NOW!

And now about our situation which is, I'm willing to bet, typical. Mom and Uncle Sonny are each descended from the oldest daughters of Thomas and Judah Farrell. You can read about them and the Farrell Project here. Mom descends from Mary Elizabeth (Farrell) House (1835-1919) and Uncle Sonny descends from her sister Catherine (Farrell) Boxwell (1838-1910). I'll show you both trees on down.

As we work through Mom's and Uncles Sonny's DNA matches we're always looking for people who match someone in this cluster of people and surnames that fan out around the Farrells. The main surnames are: Farrell, House, Hartley, and Biggerstaff.

Now the Biggerstaff surname is interesting and there's a distinct way that it's important to Mom and I. Samuel Albert House (1832-1917), the husband of Mary Elizabeth Farrell, was the illegitimate son of Isaac Biggerstaff (1798-1844). Proving this paternity is one the top items on my wish list of what I'm looking to find using DNA for genealogy. In order to accomplish that task I would have to find an undeniable DNA match to Mom who has a solid tree tracing back to an offspring of Isaac Biggerstaff through his marriage with Elizabeth Longstreath.

But there's another big problem: Samuel Albert could have also received Biggerstaff DNA from his mother. Keep reading to see how.

Now this next part is a bit sticky and complicated and I hope that the two trees below will help. Back to Uncle Sonny. The top tree for Uncle Sonny's ancestors shows the line back from James E. Boxwell, husband of Catherine Farrell. You'll notice that his mother is Dinah House, and her parents are James House and Margaret Hartley. Now look at Margaret Hartley's mother! Rebecca Biggerstaff! Which means that any of the descendants of James Boxwell and Dinah House could have Biggerstaff DNA... and in theory Uncle Sonny should too.

Uncle Sonny is a descendant of Aaron Boxwell and Dinah House.

Dinah House's grandmother was a Biggerstaff.

Now look at the this tree segment below from Mom's tree. There's Isaac Biggerstaff, presumed father of Samuel Albert House. (Are you wondering about the surname and why Samuel Albert took his mother's surname? He didn't at first and you can find him in the 1850 census listed as Samuel Biggerstaff and living in the home of his mother and step-father, Patrick Caton.)

In this tree below you'll see the biggest problem for me, and that is that Samuel Albert's father was a Biggerstaff and on his mother's side, his grandmother was a Biggerstaff. As a matter of fact, Samuel Albert's great grandmother on his mother's side was sister to his grandfather on his father's side. That's a whole big mess of Biggerstaff DNA! Is there any chance at all for me to sort it out and make a case for Isaac Biggerstaff being Samuel Albert's father using DNA?

I know, I know, I could do some Y-DNA testing with direct males descendants of SA House and Isaac Biggerstaff. I'm trying!

Here, I should mention that we've found a Biggerstaff match with Mom on AncestryDNA and he's a descendant of Isaac and Elizabeth. Nice, huh? He's Cousin Joseph and he came up with a 95% confidence rating. He's great to work with and has already shared some very useful info about local records:)

So, if Uncle Sonny has Biggerstaff  DNA he should in theory show up as a match with others who have this Biggerstaff DNA. Except that Cousin Joseph matched Mom but not Uncle Sonny. Hmmm. See, I wish Ancestry had a chromosome browser because I could use it to see right away how Cousin Joseph and Mom match and on which chromosomes.

Cousin Rich and I are scratching our heads and wondering why. Why new-to-us Cousin Joseph, the direct descendant of Isaac Biggerstaff and his wife Elizabeth Longstreth, should match Mom and not Uncle Sonny. Two answers come to mind immediately. First is that Joseph and Mom both have DNA that comes down through Isaac and no one else in our Farrell group, and I'll need a chromosome browser to answer that question. The other answer is that there is a problem with AncestryDNA's matching function. Or maybe it's both. Next time I'll talk about an issue we might have uncovered with AncestryDNA's search function.

Now do you see why I really, really want AncestryDNA to get a chromosome browser? Soon.



The URL for this post is:
http://nutsfromthefamilytree.blogspot.com/2014/05/the-thomas-and-judah-farrell-project.html