Why DNA Ethnicity Estimates Matter (and Why They Don't)

 

Those colorful break downs of your heritage can be exciting and insightful, but they've also sparked controversy. Here's a balanced look at the pros and cons:

The Concerns:

• Accuracy isn't perfect: It's true that ethnicity estimates are, well, estimates. They're based on comparing your DNA to reference populations, and the accuracy depends on factors like the size and diversity of those populations.
• Potential for division: Some worry that focusing on genetic differences could fuel ethnic conflicts. Exacerbate racism. 
• Misinterpretation and misuse: In the wrong hands, ethnicity data could be used to justify racist beliefs or discriminatory actions.
• Uneven accuracy: The accuracy of your results can be influenced by your ancestral origins. Some regions and ethnicities are better represented in databases than others.

The Counterarguments:

• Improving accuracy: Companies like AncestryDNA are constantly refining their algorithms and expanding their reference databases, leading to more precise ethnicity estimates.
• It's mostly just fun: For most people, exploring their ethnic origins is a harmless and entertaining way to connect with their heritage.
• Responsible use: The vast majority of people use their DNA results responsibly, to learn more about their family history and connect with relatives.
• Genealogical benefits: Ethnicity estimates can be a valuable tool for genealogists, helping to identify which side of the family a DNA match might be on.
• Understanding limitations: Most users are aware that ethnicity estimates are not absolute truth and should be taken with a grain of salt.

The Bottom Line:

While it's important to acknowledge the potential downsides of ethnicity estimates, it's equally important to recognize their value and the positive impact they can have. By approaching these results with a critical eye and a sense of responsibility, we can use them to enrich our understanding of ourselves, our families, and our shared human history.

How Are the New AncestryDNA Ethnicity Results Looking? Including Comparisons With First Cousins

 

The new AncestryDNA ethnicity results are looking pretty good! They're a lot better than when I first tested back in 2014. I even have French admixture now, which is a first at Ancestry. I'm glad that my French Canadian great-grandfather is now represented in my results! 

French Canadian Ancestry

I was curious about how much DNA I'd inherit from a great-great-grandfather, so I asked Gemini and ChatGPT. They both said the average is around 6.25%, but there's a slim chance I might not share any DNA with him at all. I know that's not the case for me because I have French Canadian matches and I'm in a few Quebec Settlers groups on Ancestry. It always seemed odd to be in those groups without having any French ancestry show up in my results. Now, with the update, I have 3% French! That seems about right.

Irish and German Ancestry

My Irish went down a bit, from 22% to 19%. My German went up to 23% and comes from my dad's side, which makes sense. His dad, Rudolph, was from a German-speaking Austro-Hungarian family. Ancestry says my mom is about 13% German, but Sideview says I didn't get any German from her. That doesn't seem right since I have matches on her German lines.

The best improvement is that my grandfather Rudolph Kapple is finally represented in my ethnicity estimate! It's not 100% correct, but it's close.

Comparing Cousin Results

I compared my results with my paternal first cousins (unfortunately, no maternal first cousins have tested with Ancestry). It highlights some probable errors with the estimated percentages.

Our paternal grandfather was Austro-Hungarian, with ancestors from both Austria and Hungary. He was born near Graz, Austria. When I first tested, I had NO German admixture. Now I have 23%. Here's how my German admixture compares with my cousins:

• Me: 23%
• Cousin Darryl: 7%
• Cousin Judy: 19%

Judy's 19% isn't far off from my 23%, but Darryl's 7% doesn't seem accurate. He has 13% Eastern European and 3% Balkan on our shared Kapple line, which would also represent our grandfather's heritage. He even has some Spanish on that line. If you add those up, it represents our grandfather's heritage much better. Judy also has 5% Central and Eastern European, so her totals represent our grandfather well, too. It's great to finally see our grandfather's heritage showing up in our results!

These cousins and I also share an Irish great-grandmother. Here's how our Irish results look:

• Me: 19%
• Cousin Darryl: 27%
• Cousin Judy: 35%

Sideview says our Irish comes from our shared grandmother, Dorothy Mason-Kapple. She was half Irish; her mother, Helen Mullen-Mason, was born in Ireland. We should each have around 12.5% Irish from her. Every Ancestry update brings my Irish percentage down, which is probably correct. I think they tend to overestimate Irish ancestry. Judy likely has Irish on both sides of her family, so it's not an easy comparison.

Here's a comparison of our French admixture:

• Me: 3%
• Darryl: 4%
• Judy: 0%

Since we share around 6.25% of our DNA with our great-great-grandparents, my results and Darryl's seem pretty accurate. Judy's result isn't; she should have inherited some French DNA from our great-great-grandfather.

Comparing my results

If you compare these results from 2014:

And this year's update:

You can see how far these tests have come! 

Overall Impressions

Overall, for what these tests can do, the results are pretty good! If you compare my results from 2014 and the recent update, it's clear that AncestryDNA ethnicity estimates have come a long way!

Total Shared cMs Are Important, A Chromosome Browser Is More Important

 

I  asked Gemini AI which technique of evaluating DNA matches is the best? Is total cMs shared or comparing segment data better? I agree with this chatbot that if you are serious about using DNA for genealogy both segment data and a chromosome browser are best

It is true that total cMs can be helpful when it comes to identifying close family. The new Ancestry Pro Tools Advanced Matching allows for comparison of cMs with shared matches, and matches you share with them. It also shows relationship predictions of your matches shared matches. It can also spot NPEs when total cMs are very different from what is expected. 

Total cMs isn't as helpful at all for matches beyond a couple generations. As you see below total shared cMs can vary quite a bit between first cousins, and matches we share. I share 173 cMs with a 2nd cousin 1x removed. My cousin Judy shares 248 cMs, and my cousin Darryl shares 350 cMs. My cousin Darryl shares 177 cMs more than I do with our 2nd cousin 1x removed. 

Generally I share around 1000 cMs with my first cousins. Oddly my mother shared fewer cMs with a couple 1st cousins descended from the same uncle. One first cousin shares only 534 cMs with my mother. I thought possibly there was an NPE. No, this cousin has matches on the correct lines. Another cousin, who is a half first cousin, only matches on our maternal grandmother's side and shares around 500 cMs also. Lower than expected matches need further evaluation. 

This match's brother shared only 782 cMs. Both of these cousins tested at AncestryDNA. I wonder if Timber chopped down some segments leading to these lower cM shared totals? 

My mother only shares 13 cMs with a distant cousin, but a first cousin shares 32 cMs. This demonstrates the wide range of cM totals, and the difficulty of using autosomal DNA to predict relationships. Different relatives may share more or less DNA with the same distant cousin.  

With such a wide range of relationship possibilities total shared cMs aren't that helpful. Building a chromosome map with segment data you can associate certain segments with particular families. Just looking at cMs can be helpful for close relationships, but can be a waste of time for distant relationships. 

A chromosome browser is definitely essential when trying to confirm distant matches, and solve a brickwall. I've been going through my chromosome map analyzing segment matches, and making some corrections. I can't extend it much farther at this time because fewer people test at MyHeritage and Family Tree DNA. 23andMe no longer allows you to download shared match segment data. It's difficult to get matches to upload to GEDmatch. Many just don't understand how, or what information it would provide. 

I've marked my segments, from my Genome Mate Pro segment map, with names of the ancestors who contributed them. I'm not able to visually phase my segments to recreate the segments passed down to me by my grandparents because you need siblings to do that. I'm an only child. I marked the segments I received from each grandparent using Microsoft Paint. I was thinking of using DNA Painter, but that would have taken too long. I would have had to upload segments, and mark them from scratch. 

Since AncestryDNA is unlikely to ever provide segment data my hope is 23andMe allows this data to be viewed and downloaded again, and the other companies attract more testers. Otherwise, I can't make any more progress using DNA. 

  

A Sad Message On A Photo

 

 My mother Edna Forgey-Kapple was born in Granada, Nicaragua. Her father Charles Lynn Forgey was a US Marine serving in Nicaragua. He married my grandmother Graciela Del Castillo, a Nicaraguan, in 1919. My mother was born in April 1921. 

My grandfather Charles Lynn Forgey returned to the United States without my mother and grandmother at some point after my mother's birth. I've found him referred to as Dick Forgey in Brownstown Banner newspapers. His hometown newspaper in Jackson County, Indiana. 

Here we see my grandfather mentioned in a visiting newspaper announcement published July 4th 1923. 

In August 1923 Dick Forgey is again mentioned in the Brownstown Banner. 

The fact my grandfather had been in Nicaragua hasn't been mentioned and likely hidden. My mother thought he had no plans to bring her and my grandmother to the United States. He may have wanted to start again and abandon them. 

My grandmother's father had died and he supported the family. My grandmother struggled to feed my mother after her father left. 

My grandmother had no choice but to track down my grandfather by contacting his family in Indiana. She sent the photo with the message below of my mother asking if he remembered her? 

My mother believed his father probably encouraged my grandfather to do the right thing and support his wife and child. 

In 1925 my grandfather finally did the right thing and brought his wife and my mother out to join him in California.

My grandfather did show some pride in my mother when he sent this photo to his family in Indiana. He said my mother could speak both Spanish and English. I am sorry my mother never saw this photo and the writing on the back. I had never seen it until recently. 

Was this a happy ending? It was good for my mother and grandmother to have been rescued from a difficult life in Nicaragua. My mother, however, felt she had been abandoned by her father all of her life, and it really hurt her. It saddened me to see the photo with my grandmother's plea on the back.  

My grandfather Forgey was too young and immature to marry in 1919. He was only 20 years old. I never thought about his age before. I never really did the math. I think I would have talked to my mother about his age if I had known his age at marriage to help her try to overcome sadness and feelings of abandonment. 

I feel fortunate that I always felt loved by my parents and never experienced a sense of abandonment. I'm sorry about what happened to my mother. 

New Ancestry DNA Enhanced Shared Matches Pro Tool Feature/Helpful and Meh at the same time

 

I have subscribed to Ancestry for a month in order to use the new "Enhanced Shared Matches" Pro Tool showing how many centimorgans, cMs, your matches share with each other. This has been a wonderful feature which has been available for years at 23andMe, and MyHeritage, at no additional charge. You need to have a regular Ancestry subscription in order to add Pro Tools, which is an additional $10 a month. 

My main reason for using this feature is so I can see how shared matches are related to each other on a brickwall line. I'm using this mainly with my mother's kit. I don't really care how all 30,000 of my mother's matches are related to us. I'm just interested in matches descended from George Lafayette Campbell. I've been trying to resolve a brickwall on our Campbell line for over 20 years. Our ancestor Sarah Campbell died young. She also died before the 1850 Census, so we don't have any information from when she was alive as to where she was born? She was married to Anderson Wray and had a few  children before her death. Only one Campbell was connected to the Wray family, his name was James Campbell and there is no surviving information about where he was from, or his relationship to Sarah Campbell-Wray? The children of Sarah stated on census records that she was born in Tennessee, and that is the only information about her they seemed to have. 

Looking at my mother's matches related to Campbells I noticed there were a number who descended from George Lafayette Campbell of Greene County, Tennessee. These matches don't appear to relate to us any other way. If matches are closely related to each other such as siblings, first cousins, parents and children, it could be that we aren't related through the Campbell line. We may share a different line farther back. If they don't all descend from the same children of George Lafayette Campbell the circumstantial case is stronger. If we had the underlying segment information the case would be stronger still, but AncestryDNA doesn't provide this information making the process of attempting to establish relationships slower, and more of a hassle. 

I'm gathering up the shared centimorgan information for descendants of George Lafayette Campbell and using a spreadsheet to document the shared cMs. I'm not documenting the shared cM's of every single descendant, only those sharing over 20cMs with my mother, or over 100 cMs with their shared matches, and those descended from different children of George. I want to make sure the descendants aren't closely related, or double cousins, or there wasn't endogamy in these families. 

I am discovering more close family of my matches, and noting those. Problems I'm running into include few people with trees or very limited trees. As I go through the matches I'm again looking for matches with more complete trees. One match is related through another line. It's crucial to have segment data to sort out matches who may not be related through Campbell. 

It's a slow process sorting through the information because AncestryDNA has information on so many pages, unlike MyHeritage where you can see all of the information about a match on one page. I've created notes for matches who descend from George Lafayette Campbell. I haven't been able to filter notes to bring up only notes regarding those descendants. To get to these matches I have to use the Chrome Browser search feature. Every time I return to the match list I have to search the notes again, which is another reason I'm only documenting a few descendants of George. 

When I Googled to see if there was a way to search notes I got this AI response: "As of January 2023, it's not possible to search notes on AncestryDNA, but users can use custom groups to help. However, users can find notes and comments for people in their tree on that person's profile page. Notes and comments can be found in the research tools bar to the right of the profile photo, and in the Tools menu."

I'm finding some of the matches are very closely related, such as parent child, aunts and uncles. Most matches aren't, which is good for my hypothesis. 

Before using the new Pro Tool the matches attached to this tree (below) were our main connections to the George Lafayette Campbell family. Using the shared matches tool I've found close family members of these matches that don't have trees, or a tree with a couple people on it. 

I've also looked at some close matches we didn't know how we were related to. Some of these matches have close relatives which I've found with the new tool. A couple of their matches have trees allowing me to find our relationship. We have thousands of matches and I'm not interested in determining how every match is related. This has helped with a few close matches, which is good. 

I now have 29 matches sharing descent from George Lafayette Campbell. There are many more shared matches without trees who likely share that descent too. 

This is the spreadsheet I've made to track all the shared cM totals and predicted relationships. The predicted relationships seem a little too close in some cases. I'm recording the match names across the top and side of the spreadsheet and filling in the information. There may be a better way, but this is working for me. 

Here is an example of the information I'm recording. This shows how Mary is related to our shared matches in the Campbell line. The names of the matches are on the left (I didn't include them for privacy reasons). I'll probably add which George Lafayette Campbell child each match descends from. 

It is good to know how many distinct matches you have when looking at shared matches. Sometimes people test a bunch of close relatives so your shared matches may be closely related. If they are closely related and have not extended their tree far enough back you may not be able to determine how you are related. Creating chromosome maps helps confirm relationships. You can't do that at Ancestry. 

Also, if AncestryDNA had more information about a match on one page it would save a lot of clicking back and forth.

This new Pro Tool is very helpful, but not a game changer. A chromosome browser or segment data would be the game changer we really need. 

Update: All Farmhouses Look Alike / and new Tabitha Wray information

Update from last post: I discovered looking more closely at the house I thought may have been in a photo shared on Ancestry.com, as a Moses Wray house, that it doesn't match the house in the photo. I knew it wasn't his house, he would have lived in a log house, but thought it could have sat on his land. 

The only house I could find similar to that house, using Google Street View, compared more carefully, can't be the same house in the old photo. When I looked at Google street view from another angle I could see the house didn't have 3 windows on one side. It has only one window on the side. 

Posting about my mistake on Facebook several people identified the house in the photo as being a house on Poff Lane. The house was at one time owned by a Wray family, but the original owners of the house and property were Wrights. This 1906 built house was constructed around a log cabin, likely belonging to the Wrights, sometime before the Civil War. 

Looking more carefully at the Inn it doesn't have 3 windows on the side of the house. The second floor of the house was destroyed by fire and rebuilt. I suppose windows may have been removed when the house was remodeled? An old photo taken in the 50s or 60s also shows only two windows, however I don't know which side of the house the photo was taken on? Also the slope of the land, looking at the right side of the photo is different. 

Someone would have to verify the house in the photo on Poff Lane is the same house. It may not be. Many farm houses were built just like these. The house in the photo may have been torn down, or remodeled? I think we are learning which houses are not the house in the photo. 

Old photo

What we do know is that Moses and Elizabeth Wray owned property on both sides of White Oak Bottom Creek, north of Maggotty Creek. 

Some unanswered questions about the location of their land. Is White Oak Bottom Creek the same as today's White Oak Creek? Has White Oak Creek always followed the same course? Water can always change course and people sometimes change its course. What we do know is the creek existed as early as 1755 when Jester Cox made a survey on that creek. 

Moses and Elizabeth also own land on Maggotty creek. So I'm figuring their land on White Bottom Creek was likely near where White Bottom Creek emptied into Maggotty Creek? 

New Wray Information

I've questioned whether Moses Wray's wife Elizabeth Morris could have been the mother of all 18 of his children. That seems to be more children than possible for one woman to give birth to. It is possible though. It's happened, but not frequently. My ancestor Benjamin Wray was one of Moses Wray's oldest children. So was his mother Elizabeth Morris? 

Robert Grant shared some of his research with me. His research, on the Wright family, can be viewed at FamilySearch. Sorting some of the Wrights of Southern Virginia : part III, 1809 William Wright of Franklin County, his wife Mary (Grant) Wright and his descendants Sorting some of the Wrights of Southern Virginia by Robert Grant contains information about many associated families. 

My ancestor Benjamin Wray had a sister named Tabitha Wray-Wright. 

According to Robert Grant's research Tabitha Wray-Wright's daughter married in 1792. What does this mean? It means she was likely born around 1761 or earlier. I had about 1767 as the approximate year she was born. It's very likely Elizabeth named Tabitha after her mother, her mother's name was Tabitha. My ancestor was born a few years earlier than Tabitha, but she was born much earlier than 1767. This suggests that most of Moses Wray's children were Elizabeth's. So Elizabeth Morris could be my ancestor Benjamin's mother. 

  

The information provided by Robert Grant proves there is more information out there that maybe I've missed? Eventually my questions may be answered. 

Using Parcel Maps to Find Ancestors' Land/Current Owner

 When I visited Franklin County, Virginia last August I mistakenly thought our Wray family land was located on Maggotty Creek (Maggodee Creek today). When you research the Moses Wray and Elizabeth Morris family online that creek is generally mentioned. If I read the deeds more carefully I would have noticed most of the land transactions involving Moses and Elizabeth, and their descendants, actually refer to their land being on White Oak Creek. Maggotty Creek is also mentioned because it appears their land was located in an area where White Oak Creek empties into Maggotty Creek. 

Someone on Facebook provided a link to an old photo they said was of an old Wray family house. Not old enough to be the home of Moses Wray and Elizabeth Morris, who lived during the 18th and early 19th century, but could be on their land, and could have belonged to a descendant. 

I found a similar house using Google street views from previous years. 

The photos below could be of the same house? The oldest photo, apparently, was taken decades ago. The other two are screenshots from Google Street View from 2009 and 2012. It's a little hard to compare because one of the houses has been abandoned for quite awhile. There are some differences in the sloping of the land between Google Street View and the old photo. Street View often doesn't show the sloping of land correctly. I think the gable with the window was removed when they installed a corrugated metal roof. There isn't another house in that area that looks like that house.

Sadly the old abandoned house has been torn down, and this is what the site looks like today. 

How do I confirm the Wray family owned this land and house? I tried using deeds to trace the Moses and Elizabeth Wray land forward. Deeds were not always filed, and the land was divided up so many times it's difficult to follow. It's possible the current owners of the land, where the house sat, are related to the Wrays, or they might have some information about previous land owners? Parcel maps, many are online, provide the names of the current owners of a piece of property, and sometimes recent previous owners. The name of the current owner also allows you to trace a piece of property backward using deeds or tax records. 

Since I had the address number of the property I was able to find the parcel using the Franklin County, VA - Online GIS Office. From the home page you select the Comprehensive Parcel Viewer. You can click on any parcel to find the owner, or use a name, or address. 

The Franklin County, Virginia parcel maps provide the names of the current owners of a piece of land and recent land transactions (sometimes older transactions). If there is a house on the property the size and age of the house are included. Importantly the information includes the deed book and page, plat map, and page. 

Here is the information you get when you click on a parcel.

If you click on the property card (at the bottom) you get even more information. 

Sometimes photos of the structures on a property are included. Since you can't always see houses set away from the road using Google Street View it's nice to be able to see the structures using the parcel viewer. 

Here is an example of the land transaction information. Generally fairly recent transactions are listed. In this case we see a transaction that took place many decades ago when Juanita Wray-Murray transferred land to her parents, Frank and Bertha Wray, who died a long time ago. Juanita is a distant cousin of mine. A Thurman family now owns the land and I'm likely a distant cousin of this family also. 

I'm confidant I have found the area where Moses Wray and his wife Elizabeth Morris once lived. Sorry I missed visiting that spot on my last trip. Hopefully I can return and visit the White Oak creek area. 

I've found parcel maps for many counties, and many states online, I wish they all provided the detailed information Franklin County, Virginia provides. 

Photo Fun And Confusion Apps, DPI, Sizing, Storage /Scanning Photos and Negatives

A collection of our old family cameras

In the past decade, I've found myself inheriting family photos by the hundreds. Many aren't marked. I've been using photo facial recognition software to try to identify people in the photos. I've also been looking through old boxes for negatives of our family's lost photos. I watched a Legacy Webinar, "Got Old Negatives? Scan Them With Your Phone and These 5 (Mostly) Free Apps!" by Elizabeth Swanay O'Neal, on how to use phone apps to scan negatives. These apps produce positive images of your negatives so you can get digital copies of old photos. I've been meaning for years to try to restore the lost photos, using the negatives.

Scanning Photos

Scanning old photos has been very time-consuming. It's been especially time-consuming because I've rescanned photos more than once. I should have done more research before scanning the photos. According to several experts, it's best to scan photos at 300 DPI (dots per inch) and as TIFFs. The size of your scanned photos should be at least 1280 x 720 pixels. The one scanned below was too small to enlarge even on a computer screen. I'm rescanning everything not large enough, or that is scanned as a JPEG.

 

My laptop had to be restored to factory settings due to a virus. I haven't reinstalled my scanner software yet. I decided to use my phone to make copies of some photos. I noticed the phone actually saves my photos at a larger size. I used Google Photoscan app for some of the photos. Some of the scans didn't come out sharp enough. I then tried using my phone's camera, and the results were sometimes better.

You can see below that the photo on the right is sharper than the one on the left. The one on the right was taken using my phone camera, the one on the left was scanned with Google's phone app. I think generally, if you're using a phone, just using your camera generally produces the best results.

A problem using the phone goes back to the DPI (dots per inch) debate. My Google Pixel phone saves photos at 96 DPI. The DPI recommended to make prints is at least 300 DPI. I've set my scanner to 600 DPI. It does sound like DPI is important if you decide to make prints.

From Google AI

I'm not planning on making prints at this time. I would like to save them at a good resolution to print in case family members would like prints. I plan on reloading my scanner software and using that instead of my phone.

I tried using a scanner at a library, but their scanner didn't automatically crop the photos like mine. When I cropped the photos myself, the DPI went below 300, down to 120 DPI.

I've been backing up the photos using Google Photos. I didn't realize they didn't automatically save photos at the original quality. Storage saver is the way they save the photos unless you choose original quality. The photos are compressed to save storage space. That might be fine, but if I've gone to the trouble of producing high-quality scans, I want high-quality backups.

Scanning Negatives

As I said, I watched the Legacy Webinar about scanning negatives using phone apps. I have not made a lightbox yet. A lightbox would produce the best quality prints. I've installed Kodak, Photomyne, and Filmbox negative scanning apps on my phone. I'm using Filmbox the most. That app seems to produce the best results for me.

Different size negatives I've found. The bottom 35mm negatives, on the bottom, are easiest to develop. 

Filmbox suggested using a laptop screen to scan your negatives. I have actually gotten better results using my iPad, which I can lay down flat.

This scan turned out well. The only thing I don't like is the graininess; otherwise, it's acceptable considering I no longer have the original. Better than nothing.

I took the negatives to Samy's Camera, in Pasadena, California, and had them scanned there. Most photo developing places don't scan or develop old Instamatic camera negatives. Below is the photo scan they produced.

The image produced by Leona at Samy's Camera isn't grainy and is lighter in color. The cost to scan these at Samy's Camera is $2.00 per digitized negative photo. That is the least expensive scan. The price goes up if you want them scanned larger. For that price, I received a CD with the scans, and I brought in a flash drive to also have them saved to.

Leona, at Samy's Camera, got a better scan than I did, the comparison below. My Filmbox scan is dark and grainy, on the left.

These party photos are from 1973. Some of the negative images may be deteriorating a little. Some of the photos were blurry to begin with, as you see in the photos below. The old Instamatic camera didn't take good photos of moving objects.

Without the original photos, it's hard to say whether the photos developed 50 years ago were any clearer or the colors more accurate?

In the case below, I'm not sure if this old Brownie Camera negative can produce a better image? This negative is over 60 years old, taken in May 1962. A wedding photo of my parents. I'll eventually have a professional scan it.

I love old family photos and want to make sure they are backed up and not lost so we can continue to relive happy family moments.