Nailing Down The Origin of atDNA Segments

Nailing down exactly which ancestors passed down our DNA segments is challenging for those of us with early America ancestry. It's not that our trees always collapse, mine never has, instead we share an ancestor, or an ancestral couple, or several ancestral couples. To overcome this hurdle it's important to collect as many 2nd and 3rd cousins' segments as possible. These segments are most likely to be IBD, identical by descent.

This week I was able to add a third cousin to chromosome 20 on my mother's chart. This segment can now be named Forgey/Roller after the most recent ancestral couple they share, i.e. Andrew Forgey and Anna Roller.

Marking my well documented known lines helps to eliminate these chromosome regions as areas associated with my brickwall lines.

Right now I'm examining segments on chromosome 5 and 18 as possibly relating to my Campbell family brickwall. I'm looking for the names of the parents of Sarah Campbell, who married Anderson Wray 1833 Indiana, and her place of birth in Tennessee.

I've found the common ancestors for many of our predicted high quality matches at AncestryDNA. My mother and I have had a high quality match for years at Ancestry that I can't place. The fact this match has Campbell ancestors from Tennessee is a possible promising lead for my Sarah Campbell's family. Since this person recently uploaded their raw data to Family Tree DNA I've been able to compare her with others sharing segments on chromosome 5 and 18.

I've examined the trees of matches on Chromsome 5 and 18. I've determined that segments running up to the promising Campbell matches' segments on Chromosome 18 are most likely Wray family cousins. Several also share the Campbell surname, but several don't. Since they don't overlap these segments could come from different ancestors.

 I've found more Campbell descendants who share segments on both Chromosome 5 and 18.

Chromosome 5

Chromosome 18



I have found trees posted for some of these matches, and I researched matches myself to build trees out as far as I could. Everyone sharing overlapping segments on ch. 5 and 18 shares the same ancestral couple. They are descendants of  Reuben Ellis Morton 1856-1915, and his wife Mary A. Campbell 1859-1929, of Greene County, Tennessee. These matches are all 2nd or 3rd cousins to one another. I can't find another surname on these trees that would match with the names on my tree.

Paper trails and DNA don't always match up because there can be unknown adoptions, and affairs that no one is aware of. I have a Browning line that also traces back to Greene County, Tennessee. There would have to be some sort of unknown Browning blood relationship at the Reuben Morton and Mary Campbell generation, or before, for these segments to be Browning.

 

I'm doing some research in Greene County, Tennessee records to see if I can make a connection with the Campbell's in Indiana. There was a disagreement about who George Lafayette Campbell's parents are? I was able to find the documentation for his parents being James M. Campbell and Susannah. George Lafayette Campbell was baptized in the Sinking Springs Lutheran Church in Greene County, Tennessee.

 

George Lafayette Campbell would be in the age range to be Sarah Campbell-Wray's nephew. Could his father James be her brother? Not likely since I believe she may have had a brother named James T. Campbell? Unlikely they would name two children James? It's possible James could be Sarah's cousin?

Not being completely confident that these Campbell matches actually do match us through the surname Campbell I need to find more information to confirm or eliminate this line as contributing to this DNA. If more descendants of James Campbell and Susannah match us DNA wise that would strongly suggest Sarah Campbell-Wray's family originally came from Greene County, Tennessee. William M. Campbell is thought to be James' father.  He settled in Greene County, Tennessee very early. It would be helpful if matches other than one particular line matched us. So far I can't find any other matches with trees tracing back to this family, but I will continue to check for anyone posting such a tree, or testing in the future.

I plan on continuing to search records and check DNA matches, especially on chromosomes 5 and 18, in hopes of finally solving this Campbell mystery.

These Are The Good Old Days of Genetic Genealogy/ And A Segment Puzzle Mystery?

As I'm going through Genome Mate Pro adding named segments to my segment map this week the differences in the size of the segments between my mother and I were so apparent. Many of the segments I share with matches are less than 20 cM's. My mother has many more segments over 20 cM's, and many over 30 cM's. She was born in 1921, whereas I was born in 1963. As the generations pass our genetic links to our immigrant and early American ancestors will not be visible in our DNA. Definitely test all older relatives first.

I've been mapping my segments in Genome Mate Pro. My main focus right now is on the segments shared on chromosomes 5 and 18 which relate to possible Campbell family matches. Since the only thing I know about Campbell is I have an ancestor named Sarah Campbell who married Anderson Wray in Jackson County, Indiana, possibly born in Tennessee, that's everything I know. I updated the GEDmatch data with new matches. I've been going through these segments and naming them if I can determine the ancestors they came from. This process involves using Ancestry trees. Also Googling names and email addresses to see if there is a family website.

Gedmatch is estimating the length of the segments as longer than AncestryDNA. I've filled in more of chromosome 18 adding matches besides the Campbell match. Two tested with Ancestry and overlap each other, but neither appear as shared matches. According to AncestryDNA one of these matches shares one 14.4 cM segment. GEDmatch estimates the segment is 22.4 cMs. The other match is said to share 1 segment of 12.7 cM's.  According to GEDmatch the segment length is more than twice that length at 28.3 cM's. Since Ancestry doesn't share segment data there is no way to know whether their segments overlap, according to their calculations. They do overlap according to GEDmatch.

Examining the family trees of theses matches I find that the two matches segments I've added to chromosome 18 might share more than one set of ancestors with me. These two added matches share the surnames Campbell and Wray with us. My ancestors Anderson Wray and Sarah Campbell married in 1833. On first examination it could be that these matches are also their descendants, but they aren't. Instead they are descendants of Anderson Wray's grandfather Benjamin Wray and his second wife Patsy Goode. The Campbells they descend from didn't live in Franklin County, Virginia where the Wray family lived.

 

A third match also shares DNA on chromosome 18. This match overlaps substantially with one of the other matches. Since one of them tested at Family Tree DNA and the other at AncestryDNA, and only one uploaded to GEDmatch, there is no way to tell if they match each other?

What is particularly interesting about the match at Family Tree DNA is they have ancestors who lived in Iola, Kansas. Iola is a very small town. Anderson Wray migrated with some of his children to that town. He died there. According to this match she doesn't have any relationship to the Wray family? Either this is a coincidence or there was a non paternity event? If this segment on chromosome 18 came down from Benjamin Wray then an NPE is likely. On the other hand if the match relates on the Campbell line then her tree is correct.

Last week I discovered one particularly interesting Campbell matches ancestors were from Greene County, Tennessee. Sarah Campbell-Wray was said to have been born in Tennessee. The recently added matches, who tested at AncestryDNA, have Campbell ancestors from Lee County, Virginia and Hawkins County, Tennessee. All of these places are close together in Northeast Tennessee and Southwest Virginia. My tendency is to think these segments are related to Wray and not Campbell because both share Wray ancestors?

Looking at chromosome 18 at Genome Mate Pro, after marking all of these segments, it's impossible to see where segments begin and end.

Charting the segments with Kitty's Segment Mapper we see exactly how they fit together (chart below). Not sure why there is an overlap of the blue at Genome Mate Pro? This view is very helpful. We see the Campbell match I am most interested in with segments bookending the others. I found another match who shares one of the exact segments with her. This matches has a very small tree containing no Campbell's or Wray's.

Based on this view my mother either inherited one long segment of DNA from her Great-Great Grandfather Anderson Wray, or chromosome 18 is mix from both Anderson Wray and Sarah Campbell? The book ending may suggest there is a mix of Wray and Campbell DNA on chr. 18? The Campbell match of special interest at AncestryDNA, the booking ending match, has a fairly complete tree out to around 6 and 7 generations with no Wray's on it?

The shared matches feature at Ancestry didn't give me a clue that 3 matches had something in common, facts important to me were hidden. It would have been helpful to know they share either overlapping segments, or segments that are close together. None of these matches appear to match each other even though two descend from Benjamin Wray and Patsy Goode, and both tested at AncestryDNA? They don't show up as shared matches? They do match at GEDmatch.

Doing this analysis required quite a bit of time. No one site has everything needed to effectively analyze and compare matches. I had to go back and forth between Ancestry, Family Tree DNA, 23andMe, GEDmatch, Genome Mate Pro, and Kitty's Segment Mapper. 23andMe has to best tools, but no trees. If they would add a tree feature I would recommend everyone test there.

I will continue to work with the matches on chromosome 18. I do have many more matches I need to evaluate. This kind of analysis is only possible if you have segment data. This analysis is necessary if you are comparing with distant cousins.

Brickwalls DNA and Privacy

AncestryDNA Changes

Sorry to hear that AncestryDNA plans on making changes which could hamper our ability to workout how we are related to someone. A match can now choose not to show up in results, and testers will be given the option to keep shared matches private. This is a positive development for those testing for ethnicity results who don't want to hear from matches, not so positive for the rest of us..

23andMe has managed to preserve privacy while providing needed tools like the chromosome browser and shared matches.

It's sad Ancestry has chosen to focus on ethnicity results rather than helping us to resolve our brickwalls. They've chosen to focus on the least accurate portion of the DNA results. Not providing the ability to map segments means the power of this tool will never be fully realized for their customers.

The only thing the ethnicity results can tell you with any accuracy is whether you are European, African, Middle Eastern, Native American or Asian. It can only give you vague estimates about anything more specific which actually may turn out to be completely wrong.

Working on Brickwall using DNA

After writing last week's blog post I downloaded my DNA matches from Family Tree DNA into Genome Mate Pro. Someone asked me about the DNA segment chart I included in that blog post. I was thinking it's been a couple months since I updated that database, so I needed to do that. The Family Tree DNA database has grown so slowly that updating every few months has been enough.

After completing the update to Genome Mate Pro I noticed that the Campbell match I was referring to at AncestryDNA has now transferred to Family Tree DNA. Great for me! She has a very complete tree going back about 6 generations at Ancestry. She has no tree or surnames at FTDNA so when I've been searching by the surname Campbell her name didn't come up. She transferred her kit recently.

Why is this such good news? The only information those of us researching Sarah Campbell-Wray have is that she married Anderson Wray in 1833, in Jackson County, Indiana and died there in 1848. We have a list of her children. All but two died before the 1880 US Census, when they would have stated where they thought she was born. Polly Thurman Wray-Hall stated her mother was born in Tennessee, but my ancestor Elizabeth Wray-Forgey didn't know where she was born?

With so little information, and the fact matches would be 4th cousins to my mother, probably removed by one or two generations, the segment data is crucial to making connections with DNA. I need to have a segment map which is as complete as possible in order to eliminate the possibility that we are related on another line. I have to rely on segments to give me clues. With these clues I can do further research in the actual surviving records.

Let's examine the exact segments shared with this match. Both my mother and I share the same segments and same and number of cM's with this match.

Chromosome 5, 28.05 cM's , SNP's 6100

Chromosome 18, 5.69 cM's , SNP's 1700

Chromosome 18, 22.27 cM's , SNP's 2443

We share a total of 56.01 cM's according to Family Tree DNA on 3 chromosomes. According to AncestryDNA we share 43 cM's on 3 segments.

Comparing these segments to others in the same places on the chromosomes it's looking promising that these segments do indeed relate to Campbell. They don't match up with any known Roller matches. They don't match up with any other surname segments.

Doing further comparison with shared matches at Family Tree DNA more good news! Half a dozen shared matches also have the surname Campbell listed in the surnames. As I stated before Sarah Campbell was said to have been born in Tennessee. Several of those who have Campbell listed have Georgia/Alabama ancestors. We know that there was migration heavy migration from Tennessee farther south. I was blown away when I saw two of these matches had Campbell ancestors who were from Greene County, Tennessee, which matches the place of origin of our AncestryDNA match. There was some migration from Greene County, Tennessee to Lawrence and Jackson County, Indiana. My Browning family migrated to these counties from Greene County, Tennessee.

Greene County sounds like a promising place of birth for Sarah Campbell? I will comb those records for any connections.

What I'm getting out of the DNA testing is confirmation of already well documented lines. DNA is matching the paper trail meaning there were no unknown adoptions etc. When it comes to my brickwalls more work is required. All of my brickwalls are at or beyond the 4th generation. The chances of matching a 4th cousin and beyond is far less likely. You have a 90% chance of sharing DNA with a 3rd cousin, but that drops to 50% with a 4th cousin and only 10% with a 5th cousin. Building a segment map using 2nd and 3rd cousins is the most effective way to workout those more distant matches. These large segments are more likely to be IBD, giving a firm foundation for those smaller segments.

I'm going to focus on building my segment map in hopes of narrowing down who is related on my brickwall Campbell line.