I wanted to evaluate the accuracy of Ancestry’s ThruLines suggested Potential Ancestors when compared with a tree I know is accurate. I conducted an experiment where I created a small tree on Ancestry for a DNA tester that included only the first two generations, meaning grandparents and great-grandparents.

This gave Ancestry enough data to work with and means that for the upstream ancestors, Ancestry’s ThruLines suggested specific people as ancestors.

How well did Ancestry do? Are the Potential Ancestors suggested by Ancestry accurate? How do they make those suggestions anyway? Are they useful?

I do have a second, completely separate, full tree connected to my other DNA test, and I do know who those ancestors are, or, in some cases, I know who they aren’t. I’ve had the privilege of working intensively on my genealogy for decades, so I can easily compare what is known and proven, or what has been disproven, to Ancestry’s suggested Potential Ancestors.

We’ll start with the great-grandparents’ generation, but first, let’s talk about how ThruLines works. I’ve previously written about ThruLines here and here.

How ThruLines Works

ThruLines is a tool for people who have taken an AncestryDNA test and who link themselves to their position on their tree. Linking is a critical step. If you don’t link the DNA test to the proper profile, the tester won’t have ThruLines. I provided step-by-step instructions, here.

I want to emphasize this again, ThruLines is a TOOL, not an answer. It may or may not be accurate and it’s entirely UP TO YOU to take that hint, run with it, and verify or disprove. Ancestry is providing you with a hint.

Essentially, the more ancestors that you provide to Ancestry, generally, the better they can do when suggesting additional Potential Ancestors. They do need something to work with. I wrote about that in the article Optimizing Your Tree at Ancestry for More Hints and DNA ThruLines.

If you don’t provide at least your parents and at least your grandparents in a tree, it’s unlikely that Ancestry will be able to provide Potential Ancestors for you.

I added two generations above the parents in this experiment in order to provide Ancestry with a significant “hook” to latch onto to connect with:

• Other DNA testers who match the tester AND
• Other people’s trees, whether the tree-owners have tested their DNA or not

So yes, to be clear, Ancestry DOES:

• Use the trees of other people whose DNA you match AND have the same ancestors in their tree
• Along with the trees of people you don’t match (or who haven’t DNA tested,) to propose ancestors for you

ThruLines only reaches back to ancestors within 7 generations, meaning the ancestor is the tester’s 5^th great-grandparent or closer.

Most suggested Potential Ancestors in ThruLines have descendants who have tested and are DNA matches to you, but not necessarily all.

On your tree itself, the ThruLines “3 people” icon shows on the ancestors that have Thrulines.

Looking at this graphic of my tree, you can see that ThruLines ends at the 7^th generation, but Potential Ancestors continue to be suggested beyond 7 generations. Note generation 9, below, which is beyond ThruLines but has Potential Ancestors suggested based entirely on other people’s trees.

ThruLines stops at 7 generations, but Potential Ancestor suggestions do not.

In the above example, in generation 7, Michael McDowell (1720-1755) is a known ancestor and has a ThruLine, but his wife is unknown. Ancestry has suggested a Potential Mother for Michael McDowell (1747-1840) who is also the spouse of Michael McDowell (1720-1755).

Here’s the ThruLines suggestion for Michael McDowell’s wife.

Ironically, there are no DNA matches for either Michael or Eleanor. However, there are DNA matches for their child who clearly descends from Michael. This may be an example of a situation where the other testers are beyond the 7^th generation, so they don’t show as matches for our tester in Michael’s generation. The other possibility, of course, is a glitch in ThruLines.

(For those familiar with the Michael McDowell (1720-1755) lineage, Eleanor is his mother, not his wife. His wife is unknown, so this Potential Ancestor is incorrect.)

Potential Ancestors Without DNA Matches

A person may still be suggested as a Potential Ancestor even without any DNA matches.

I have seen situations where a parent has DNA matches to several ThruLine ancestors, but their child has the same suggested ancestor with zero DNA matches listed because the child and the match are one generation too far removed to be listed as a DNA match on ThruLines.

Yet, if you search the child’s match list for the individual listed as a DNA match to their parent through that ancestor, that match is also on the child’s match list.

In the chart that follows, you can see that ancestors in the midrange of generations have many DNA matches, but as you approach the 7^th generation, the number of matches drops significantly, and some even have zero. That’s because both people of a match pair have to be within the generational boundary for ThruLines to list them as matches.

In some cases, the ancestor is not suggested for the child in ThruLines because the ancestor is the 6th great-grandparent of the child. If you look directly at the child’s tree, the Potential Ancestor may be suggested there.

Points to Remember

• The difference between ThruLines and Potential Ancestors is that Potential Ancestors are still suggested beyond the hard 7 generation or 5 GG boundary for ThruLines.
• ThruLines may suggest Potential Ancestors with or without DNA matches.
• Potential Ancestors, either within or beyond ThruLines must connect to someone in your tree, or another Potential Ancestor or ancestors who connect to someone in your tree.

Incorrect Ancestors and Discrepancies

An incorrect ancestor can be listed in multiple people’s trees, and Ancestry will suggest that incorrect ancestor for you based on the associated trees. At one point, I did a survey of the number of people who had the incorrect Virginia wife listed for my ancestor, Abraham Estes, and the first 150 trees I viewed had the wrong wife. We have church record proof of her death in England before his children were born by his colonial Virginia wife. Garbage in, garbage out.

That doesn’t mean those trees aren’t useful. In some cases, the information “saved” to that person in those incorrect trees shows you exactly what is out there and can’t be correct. For example, if there is a death record and burial for someone, they can’t also be alive 50 years later in another location. Or someone born in 1780 can’t have been a Revolutionary War veteran. Sometimes you’ll discover same name confusion, or multiple people who have been conflated into one. Other times, you may actually find valid hints for your own ancestor misplaced in someone else’s tree. Always evaluate.

You “should” have the same number of matches to the man and woman of a couple if neither of them had descendants with another partner, but sometimes that doesn’t happen. I would presume that’s due to tree discrepancies among your matches or other trees on Ancestry.

If the same ancestor is listed with multiple name spellings or similar differences, I have no idea how Ancestry determines which version to present to you as a Potential Ancestor. That’s why ThruLines are hints. Ancestry does show you the various trees they utilized and allows you to peruse them for hints for that suggested ancestor.

Just click on the Evaluate button. Unfortunately, neither of these trees have any records for this ancestor.

If you click on the tree, you are then given the opportunity to add Eleanor (meaning the potential ancestor) to your tree from their tree.

I STRONGLY, STRONGLY suggest that you DO NOT do this. By adding information directly from other people’s trees, you’re introducing any errors from their tree into your tree as well.

If you click through to their tree, you’ll often find that they used someone else’s tree as their “source,” so misinformation propagates easily. Seeing “Ancestry Family Trees” as a source, especially in multiple records, provides you with an idea of the research style of that tree owner. This also conveys the message to less-experienced researchers that copy/pasting from other trees is a valid source.

Use this information provided as hints and do your own research and evaluation.

Where Do Potential Ancestors Come From?

Let’s view an example of an incorrect Potential Ancestor suggestion and proof-steps you can utilize to help validate or potentially disprove the suggestion.

We know that George Middleton Clarkston/Clarkson is NOT the father of James Lee Clarkson based on Y-DNA testing where the descendants of the two men not only don’t match, they have a completely different haplogroup. They do not share a common paternal ancestor. Furthermore, proven descendant groups of both men do not have autosomal DNA matches.

However, George Middleton Clarkson is suggested as a Potential Ancestor in ThruLines as the father of James Lee Clarkson.

Mousing over the ThruLines placard shows 98 DNA matches to other people who claim descent from George Middleton Clarkson. How is it possible to have 98 matches with descendants of George Middleton Clarkson, yet he’s not my ancestor?

Many people just see that “98,” which is a high number and think, “well, of course he’s my ancestor, otherwise, I wouldn’t match all those descendants.” It’s not that simple or straightforward though. It’s certainly possible to all be wrong together, especially if you’re dealing with long-held assumptions in the genealogy community and trees copies from other people’s trees for decades.

To view the ThruLine detail for George Middleton Clarkson, just click on the placard.

The ThruLine for George Middleton Clarkson has three attributed children with DNA matches. Let’s evaluate.

• ThruLines Child 1 is my own James Lee Clarkson that has been erroneously attached to George Middleton Clarkson. However, the Y-DNA of the three various lines, above, does not match. That erroneous connection alone counts for 80 of those 98 matches. If all of those people who match me do descend from our common ancestor, James, those matches all make sense.

According to early histories, James Lee Clarkson was believed to be George’s son based on geographic proximity between the state of Franklin in eastern Tennessee and Russell County, Virginia, but then came DNA testing which said otherwise.

This DNA grouping from the Clarkson/Claxton DNA Project at FamilyTreeDNA shows that the men, above, which includes descendants of James Lee Claxton/Clarkson, all match each other.

• ThruLines Child 2 is Thomas Clarkston who has 17 DNA matches through 7 of his children.

By clicking on the green evaluate button for Thomas, we see that two of the DNA related trees have records, but three do not.

The first tree is quite interesting for a number of reasons.

1. Thomas Clarkson is found in Lee County, VA, in relatively close proximity to where James Lee Clarkson is first found in Russell County, VA as an adult in 1795.
2. There is no actual documentation to connect Thomas Clarkson with George Middleton Clarkson who was hung in 1787 in the lost State of Franklin, Tennessee, now Washington and Greene Counties in Tennessee. It has been “accepted” for years that Thomas descends from George Middleton based on information reportedly passed down within that family long before the internet.

The Claxton/Clarkson DNA Project at FamilyTreeDNA shows the Thomas lineage. This lineage reaches back into England based on Y-DNA matches – a huge and important hint for the Thomas descendants that they won’t be able to obtain anyplace else.

Note that Thomas’s Y-DNA does not match that of James Lee Clarkson/Claxton which means these people must match me through a different line. That’s not surprising given that many of the families of this region intermarried for generations.

• ThruLines Child 3 is David Claxton, who has one DNA match, so let’s look at that by clicking on the green evaluate button.

You’ll see that this ancestor through David Claxton was recommended based on:

• One DNA match with a tree with 0 source records, and
• Zero Ancestry member trees of people whose DNA I don’t match, or that haven’t DNA tested

Checking this tree shows no sources for the following generations either, so I have no way to evaluate the accurace of the tree.

However, I did track his descendants for a generation or so and found them in Wilson County, TN, which allowed me to find them in the Clarkson/Claxton Y DNA Project at FamilyTreeDNA.

In the Clarkson/Claxton DNA project, we see that this David Claxton of Wilson County, TN is in a third DNA group that does not match either the James Lee Claxton or the Thomas Claxton line.

Furthermore, look at the hints for the descendants of David Claxton based on the Y-DNA matches. This link appears to reach back to a Clayton in Kirkington, Yorkshire.

ThruLines Conflation

In this case, three men of similar or the same surnames were cobbled together as sons of George Middleton Clarkson where clearly, based on Y-DNA testing, those three men are not related to each other paternally and do not share a common paternal ancestor. They cannot all three be descendants of George Middleton Clarkson.

It’s amazing how much is missed and erroneously inferred by NOT testing Y-DNA. In very short order, we just proved that the ThruLine that connected all three of these men to George Middleton Clarkson as their ancestor is inaccurate.

In defense of Ancestry, they simply used user-submitted erroneous trees – but you have it within YOUR power to search further, and to utilize Y-DNA or mitochondrial DNA testing for additional clarification. This Clarkson/Claxton information was freely available, publicly, by just checking.

You can find surname or other projects at FamilyTreeDNA, by scrolling down, here, or simply google “<surname you seek> DNA Project.”

How Can These People All Match the Tester?

If we know that the male Claxton/Clarkson line is not the link between these matches, then why and how do these people all DNA match the tester? That’s a great question.

It’s possible that:

• They match the tester through a different ancestor
• There has been a genetic disconnect in the Claxton/Clarkson line and the match is through the mother, not the Claxton/Clarkson male
• Some of the other testers’ genealogy is in error by including George Middleton Clarkson in their trees
• People accept the George Middleton Clarkson suggestion, adding him to their tree, propagating erroneous information
• The descendants of James Lee Clarkson/Claxton match because he is their common ancestor, but connecting him to George Middleton Clarkson is erroneous
• The 15 cM match (and potentially others) is identical by chance
• The Y-DNA disproved this possibility in this case. In other cases, the matches could have been from the same biological Clarkson/Claxton line, but the testers have their ancestor incorrectly attached to George Middleton Clarkson/Claxton. In this case, we can’t say which of David Claxton, James Lee Claxton and/or Thomas Claxton are or are not individually erroneously connected to George Middleton Clarkson, but we know for a fact that David’s, James’ and Thomas’s descendant’s Y-DNA does not match each other, so they can’t all three be descendants of George Middleton Clarkston. Furthermore, there is no solid evidence that ANY of these three men are his descendant. We know that these three men do not share a common direct paternal ancestor.

I recommend for every male line that you check the relevant Y-DNA project at FamilyTreeDNA and see if the information there confirms or conflicts with a suggested ancestor, or if a descendant hasn’t yet tested. I also STRONGLY recommend that a male in the relevant surname line that carries that surname be asked to test in order to verify the lineage.

ThruLine Ranking

I’m going to rank Ancestry’s suggested Potential Ancestors by awarding points for accuracy on their Potential Ancestor ThruLines suggestions and subtracting points for incorrect Potential Ancestor suggestions. This chart is at the end with links to my 52 Ancestor’s articles for those ancestors.

OK, let’s take a look, beginning with the great-grandparent generation.

Great-Grandparents

I entered all of these ancestors and they are connected to their children, the tester’s grandparents. They are not connected to their parents for purposes of this article, although I do know who the parents are, so let’s see how Ancestry does making Potential Ancestor suggestions through ThruLines.

Ancestors (above example) that are NOT framed by a dotted line and who are NOT labeled as a “Potential Ancestor” have been connected in their tree by the DNA tester, meaning you.

The next generations, below, are all framed by dotted lines, meaning they are Potential Ancestor suggestions provided by Ancestry. Potential Ancestors are always clearly marked with the green bar.

Eight 2^nd Great Grandparents

In this generation, because I have not connected them, Ancestry has suggested Potential Ancestors for all sixteen 2X Great-Grandparents.

I’ve provided gold stars for the correct ancestor information meaning both the name and the birth and death date within a year or a decade when they died between census years.

Of these 16, three are completely accurate and the rest were at least partially accurate.

I repeated this process for each one of the suggested Potential Ancestors in the 3^rd, 4^th and 5^th great grandparent categories as well, completing a ranking chart as I went.

Ranking Chart

I’ve ranked Ancestry’s accuracy in their Potential Ancestor recommendations.

• +2 points means the name AND birth and death years are accurate within a year or decade if they died within a census boundary
• +1 point means that EITHER the name OR the birth and death dates are (mostly) accurate, but not both
• 0 means uncertain, so neither positive or negative
• -1 point means that NEITHER the name NOR birth and death dates are accurate but it’s clear that this is meant to be the correct person. In other words, with some work, this hint could point you in the right direction, but in and of itself, it is inaccurate.
• -2 means that the person suggested is the wrong person

I’ve been generous where there was some question. I’ve linked these ancestors where I’ve written their 52 Ancestors stories. [LNU] means last name unknown. It’s worth noting that one of the trees Ancestry has available to utilize for Potential Ancestors is my own accurate tree with many source documents for my ancestors.

#	Generation	Ancestry Name & Birth/Death Years	Correct Name & Birth/Death Years	# Matches	Points Awarded	Y or mtDNA Confirmed
1	2nd GGP	John R. Estes 1788-1885	John. R. Estes 1787-1885	110	2	Yes
2	2nd GGP	Nancy Ann Moore 1789-1865	Ann Moore or Nancy Ann Moore c1785-1860/1870	112	1	Need mtDNA through all females
3	2nd GGP	Lazarus Dotson 1785-1861	Lazarus Dodson 1795-1861	46	-1	Yes
4	2nd GGP	Elizabeth Campbell 1802-1842	Elizabeth Campbell c 1802-1827/1830	46	1	Yes
5	2nd GGP	Elijah R. Vannoy 1782-1850	Elijah Vannoy 1784-1850s	82	-1	Yes
6	2nd GGP	Rebecca Lois McNeil 1781-1839	Lois McNiel c1786-c1830s	81	-1	Yes
7	2nd GGP	William Crumley ?-1859	William Crumley 1788-1859	97	1	Yes
8	2nd GGP	Lydia Brown Crumley 1796-1847	Lydia Brown c1781-1830/1840	112	-1	Yes
9	2nd GGP	Henry Bolton 1741-1846	Henry Frederick Bolton 1762-1846	152	-1	Yes
10	2nd GGP	Nancy Mann 1777-1841	Nancy Mann c1780-1841	134	1	Yes
11	2nd GGP	William Herrel 1803-1859	William Harrell/Herrell c1790-1859	31	1	Yes
12	2nd GGP	Mary McDowell 1785-1871	Mary McDowell 1785-after 1872	45	2	Yes
13	2nd GGP	Fairwick Clarkson 1800-1874	Fairwix/Fairwick Clarkson/Claxton 1799/1800-1874	82	2	Yes
14	2nd GGP	Agnes Sander Muncy 1803-1880	Agnes Muncy 1803-after 1880	106	1	Yes
15	2nd GGP	Thomas Charles Speak 1805-1843	Charles Speak 1804/1805-1840/1850	60	1	Yes
16	2nd GGP	Ann McKee 1805-1860	Ann McKee 1804/1805-1840/1850	60	1	Yes
17	3rd GGP	George M. Estes 1763-1859	George Estes 1763-1859	76	1	Yes
18	3rd GGP	Mary C. Younger 1766-1850	Mary Younger c1766-1820/1830	75	-1	Yes
19	3rd GGP	William Moore 1756-1810	William Moore 1750-1826	72	1	Yes
20	3rd GGP	Susannah Harwell 1748-1795	Lucy [LNU] 1754-1832	69	-2	Need Lucy’s mtDNA through all females
21	3rd GGP	Lazarous Dotson 1760-1826	Lazarus Dodson 1760-1826	42	1	Yes
22	3rd GGP	Janet Jane Campbell 1762-1826	Jane [LNU] c1760-1830/1840	38	-2	Need mtDNA through all females
23	3rd GGP	John Campbell 1772-1836	John Campbell c1772-1838	65	1	Yes
24	3rd GGP	Jane Dobkins 1780-1860	Jane Dobkins c1780-c1860	22	2	Yes
25	3rd GGP	Francis Vanoy/Vannoy 1746-1822	Daniel Vannoy 1752-after 1794	76	-2	Yes
26	3rd GGP	Millicent “Millie” Henderson 1755-1822	Sarah Hickerson 1752/1760-before 1820	76	-2	Need mtDNA through all females
27	3rd GGP	William McNeil/McNeal 1760-1830	William McNiel c1760-c1817	116	1	Yes
28	3rd GGP	Elizabeth Shepherd McNeil 1766-1820	Elizabeth Shepherd 1766-1830/1840	115	-1	Yes
29	3rd GGP	William Crumley 1767-1837	William Crumley c1767-c1839	59	1	Yes
30	3rd GGP	Hannah Hanner “Hammer” 1770-1814	unknown	60	-2	Have her mtDNA
31	3rd GGP	Jotham Sylvanis Brown 1765-1859	Jotham Brown c1740-c1799	100	-2	Yes
32	3rd GGP	Ruth Johnston Brown	Phoebe Cole 1747-1802	97	-2	Incorrect person but have correct mtDNA
33	3rd GGP	Henry Bolton 1720-1757	Henry Bolton 1729-1765	88	1	Yes
34	3rd GGP	Sarah Corry 1729-1797	Sarah Corry 1729-1797	80	2	Need mtDNA through all females
35	3rd GGP	Robert James Mann 1753-1801	James Mann 1745-?	77	-1	Need Y-DNA
36	3rd GGP	Mary Jane Wilson 1760-1801	Mary Brittain Cantrell c1755-?	80	-2	Incorrect but have correct mtDNA
37	3rd GGP	John Herrell 1761-1829	John Harrold c1750-1825	19	-1	Yes
38	3rd GGP	Hallie	Mary [LNU] c1750-1826	18	-2	Need mtDNA through all females
39	3rd GGP	Michael McDowell-McDaniel 1737-1834	Michael McDowell c1747–1840	25	-2	Yes
40	3rd GGP	Sarah Isabel “Liza” Hall	Isabel [LNU] c1753-1840/1850	27	-2	Need mtDNA through all females
41	3rd GGP	James Lee Clarkson 1775-1815	James Lee Clarkson c1775-1815	170	2	Yes
42	3rd GGP	Sarah Helloms Cook 1775-1863	Sarah Cook 1775-1863	188	1	Yes
43	3rd GGP	Samuel Munsey-Muncy 1767-1830	Samuel Muncy after 1755-before 1820	108	1	Yes
44	3rd GGP	Anne W. Workman 1768-1830	Anne Nancy Workman 1760/1761-after 1860	107	-1	Yes
45	3rd GGP	Rev. Nicholas Speak 1782-1852	Nicholas Speak/Speaks 1782-1852	93	2	Yes
46	3rd GGP	Sarah Faires Speak 1782-1865	Sarah Faires 1786-1865	93	-1	Yes
47	3rd GGP	Andrew McKee 1760-1814	Andrew McKee c1760-1814	86	2	Yes
48	3rd GGP	Elizabeth 1765-1839	Elizabeth [LNU] c1767-1838	88	2	Yes
49	4th GGP	Moses Estes 1742-1815	Moses Estes c1742-1813	27	1	Yes
50	4th GGP	Luremia Susannah Combes 1747-1815	Luremia Combs c1740-c1820	33	-1	Need mtDNA through all females
51	4th GGP	Marcus Younger 1735-1816	Marcus Younger 1730/1740-1816	30	2	Yes
52	4th GGP	Susanna Hart* 1725-1806	Susanna [possibly] Hart c1740-before 1805	26	-1	Yes
53	4th GGP	William Moore 1725-1757	James Moore c1718-c1798	25	-2	Yes
54	4th GGP	Margaret Hudspeth 1725-1808	Mary Rice c1723-c1778/1781	26	-2	Need Mary Rice mtDNA through all females
55	4th GGP	Samuel “Little Sam” Harwell 1716-1793	Incorrect	36	-2
56	4th GGP	Abigail Anne Jackson 1712-1793	Incorrect	33	-2
57	4th GGP	Rawleigh “Rolly” Dodson 1730-1793	Raleigh Dodson 1730-c1794	19	2	Yes
58	4th GGP	Elizabeth Mary Booth 1728-1793	Mary [LNU] c1730-1807/1808	27	-2	Need Mary’s mtDNA through all females
59	4th GGP	Nancy Ann Steele 1728-1836	Unknown mother of Jane [LNU], wife of Lazarus Dodson	16	-2	Need Jane’s mtDNA through all females
60	4th GGP	James Campbell 1742-1931	Charles Campbell c1750-c1825	28	-2	Y DNA confirmed NOT this line
61	4th GGP	Letitia Allison 1759-1844	Incorrect	31	-2
62	4th GGP	Jacob Dobkins 1750-1833	Jacob Dobkins 1751-1835	91	1	Yes
63	4th GGP	Dorcas (Darcas) Johnson 1750-1831	Darcus Johnson c1750-c1835	92	2	Yes
64	4th GGP	John Francis Vannoy 1719-1778	John Francis Vannoy 1719-1778	47	2	Yes
65	4th GGP	Susannah Baker Anderson 1720-1816	Susannah Anderson c1721-c1816	59	2	Need mtDNA through all females
66	4th GGP	Thomas Hildreth Henderson 1736-1806	Charles Hickerson c1725-before 1793	37	-2	Have Hickerson Y-DNA
67	4th GGP	Mary Frances “Frankie” McIntire 1735-1811	Mary Lytle c1730-before 1794	37	-2	Need mtDNA from all females
68	4th GGP	Rev. George W. McNeil 1720-1805	George McNiel c1720-1805	143	1	Yes
69	4th GGP	Mary Sarah Coates 1732-1782	Sarah/Sallie or Mary [maybe] Coates c1740-1782/1787	139	1	Need mtDNA through all females
70	4th GGP	John James Sheppard Shepherd 1734-1810	Robert Shepherd 1739-1817	136	-2	Have Shepherd Y-DNA
71	4th GGP	Sarah Ann Rash 1732-1810	Sarah Rash 1748-1829	178	-1	Yes
72	4th GGP	John Crumbley 1737-1794	William Crumley 1736-1793	77	-2	Have Crumley Y-DNA
73	4th GGP	Hannah Mercer 1742-1774	Hannah Mercer c1740-c1773	73	2	Yes
74	4th GGP	John Hanner (Hainer)	Incorrect	19	-2
75	4th GGP	Jotham Brown 1740-1799	Incorrect	183	-2	Have Brown Y-DNA
76	4th GGP	Phoebe Ellen Johnston 1742-1810	Incorrect	182	-2
77	4th GGP	Moses Johnston 1746-1828	Incorrect	45	-2
78	4th GGP	Eleanor Havis 1753-1837	Incorrect	47	-2
79	4th GGP	Henry Boulton 1693-1737	John Bolton before 1693-after 1729	23	-2	Have Bolton Y-DNA
80	4th GGP	Elizabeth Bryan 1658-1742	Elizabeth Goaring 1795-1729	22	-2	Need mtDNA through all females
81	4th GGP	Thomas Curry (Corry) 1705-1729	Thomas Curry 1705-1729	25	2	Need Curry Y-DNA
82	4th GGP	Monique “Moniky” Curry 1704-1729	Monique Demazares 1705-1729	25	1	Need mtDNA through all females
83	4th GGP	Robert James Mann 1740-1787	John Mann 1725-1774	26	-2	Need Mann Y-DNA
84	4th GGP	Sarah Susannah McCloskey 1716-1797	Frances Carpenter 1728-1833	28	-2	Need mtDNA through all females
85	4th GGP	Benjamin “Col. Ben” Colonel Wilson 1733-1814	Incorrect	28	-2
86	4th GGP	Mary Ann Seay 1735-1814	Incorrect	29	-2
87	4th GGP	John Hugh McDowell 1695-1742	Michael McDowell c1720-after 1755	7	-2	Incorrect but have correct Y-DNA McDowell Y-DNA
88	4th GGP	Mary Magdalena Woods 1705-1800	Incorrect	8	-2
89	4th GGP	Ebenezer Hall 1721-1801	Incorrect	6	-2
90	4th GGP	Dorcas Abbott Hall 1728-1797	Incorrect	6	-2
91	4th GGP	George Middleton Clarkston/Clarkson 1745-1787	Incorrect	98	-2	Incorrect but have correct Clarkson Y-DNA
92	4th GGP	Catherine Middleton 1764-1855	Incorrect	94	-2
93	4th GGP	William Henry Cook 1750-1920	Joel Cook before 1755 – ?	83	-2	Need Cook Y-DNA
94	4th GGP	Elizabeth Wall 1747-1826	Alcy [LNU] c 1755-?	91	-2	Yes
95	4th GGP	Obediah Samuel Muncy 1735-1806	Samuel Muncy 1740-1799	33	-1	Yes
96	4th GGP	UFN Obediah Muncy wife Unknowen (sic) 1728-1843	Agnes Craven 1745-1811	27	-2	Need Agnes Craven Need mtDNA through all females
97	4th GGP	Joseph Workman 1732-1813	Joseph Workman c1736-c1813	64	2	Yes
98	4th GGP	Phoebe McRay McMahon 1745-1826	Phoebe McMahon c1741-after 1815	64	1	Yes
99	4th GGP	Charles Beckworth Speake/Speaks 1741-1794	Charles Speake c1731-1794	47	1	Yes
100	4th GGP	Jane Connor 1742-1789	Incorrect, unknown first wife	40	-2	Need mtDNA through all females
101	4th GGP	Gideon Farris 1748-1818	Gideon Faires before 1749-1821	54	-1	Yes
102	4th GGP	Sarah Elizabeth McSpadden 1745-1821	Sarah McSpadden c1745-c1820	55	1	Yes
103	4th GGP	Hugh McKee 1720-1795	Unknown	34	-2
104	4th GGP	Mary Nesbit 1732-1795	Unknown	35	-2
105	4th GGP	Private (sic)	Unknown father of Elizabeth, wife of Andrew McKee	35	-2
106	4th GGP	Anna Elizabeth Carney [wife of “private”]	Incorrect	35	-2
107	5th GGP	Moses Estes 1711-1788	Moses Estes 1711-1787	13	2	Yes
108	5th GGP	Elizabeth Jones “Betty” Webb 1718-1782	Elizabeth [LNU] 1715/1720-1772/1782	5	-2	No known daughters
109	5th GGP	George W. Combs 1714-1798	John Combs 1705-1762	6	-2	Need Combs Y-DNA
110	5th GGP	Phebe Wade ?-1830	Incorrect	6	-2	Need mtDNA of John Combs first wife through all females
111	5th GGP	Sarah Ferguson 1700-1781	Incorrect	3	-2
112	5th GGP	Anthony Hart 1700-?	Possibly Anthony Hart but no evidence	3	0
113	5th GGP	Charles Rev. Moore 1685-1734	Incorrect	4	-2
114	5th GGP	Mary Margaret Barry Moore 1690-1748	Incorrect	4	-2
115	5th GGP	Ralph Hudspeth II* 1690-1776	Incorrect	9	-2
116	5th GGP	Mary Carter 1699-1737	Incorrect	3	-2
117	5th GGP	Samuel Harwell 1674-1767	Incorrect	3	-2
118	5th GGP	Mary Ann Coleman*8th Ggm (sic) 1678-1723	incorrect	6	-2
119	5th GGP	Ambrose (Sar) Jackson 1695-1745	Incorrect	6	-2
120	5th GGP	Anne Amy Wyche 1692-1765	Incorrect	6	-2
121	5th GGP	George E Dodson (DNA) (sic) 1702-1770	George Dodson 1702-after 1756	23	-1	Yes
122	5th GGP	Margaret Dogett Dagord 1708-1770	Margaret Dagord 1708-?	24	1	Need mtDNA through all females
123	5th GGP	James Booth 1700-1741	Incorrect	4	-2
124	5th GGP	Frances Dale Booth (15great aunt) (sic) 1688-1777	Incorrect	3	-2
125	5th GGP	Samuel Scurlock Steele 1709-1790	Incorrect	2	-2
126	5th GGP	Robert R. Campbell 1718-1810	Incorrect	34	-2
127	5th GGP	Lady: Letitia Crockett 1719-1760	Incorrect	8	-2
128	5th GGP	John A. Dobkins 1717-1783	John Dobkins c1710-c1788	20	1	Yes
129	5th GGP	Mary Elizabeth Betty Moore 1739-1815	Elizabeth [LNU] c1711-?	20	-2	Need mtDNA through all females
130	5th GGP	Peter Johnson 1715-1796	Peter Johnson/Johnston c1720-c1794	0	1	Yes
131	5th GGP	Mary Polly Phillips 1729-1790	Mary Polly Phillips c1726-?	1	2	Need mtDNA through all females
132	5th GGP	Francis Janzen Vannoy Van Noy 1688-1774	Francis Vannoy 1688-1774	8	1	Yes
133	5th GGP	Rebecca Anna Catherine Anderson 1698-1785	Rebecca Annahh Andriesen/ Anderson 1697-1727	13	-1	Need mtDNA through all females
134	5th GGP	Cornelius Anderson (Andriessen) 1670-1724	Kornelis Andriesen 1670-1724	5	2	Yes
135	5th GGP	Annetje Annah Opdyck 1670-1746	Annetje Opdyck c1675-after 1746	5	2	Need mtDNA through all females
136	5th GGP	Thomas Hildret Henderson 1715-1794	Incorrect	3	-2
137	5th GGP	Mary Frisby 1709-1794	Incorrect	3	-2
138	5th GGP	Alexander (Alex) McEntire 1707-1802	Incorrect	12	-2
139	5th GGP	Hannah Janet McPherson 1711-1792	Incorrect	15	-2
140	5th GGP	Thomas James McNeil 1699-1803	Incorrect	25	-2
141	5th GGP	Mary Hannah Parsons 1697-1784	Incorrect	27	-2
142	5th GGP	John Coates 1699-1732	Incorrect	21	-2
143	5th GGP	Sarah Ann Titcombe 1710-1732	Incorrect	22	-2
144	5th GGP	George Sheppard, Shepherd 1716-1751	George Shepherd c1700-1751	42	1	Have Shepherd Y-DNA
145	5th GGP	Elizabeth Mary Angelicke Day (Daye) 1699-?	Elizabeth Mary Angelica Daye 1699-after 1750	41	1	Need mtDNA through all females
146	5th GGP	Joseph Rash 1722-1776	Joseph Rash before 1728-c1767	36	1	Yes
147	5th GGP	Mary Warren 1726-1792	Mary Warren 1726-?	36	1	Yes
148	5th GGP	James L Crumley/Cromley 1712-1784	James Crumley c1711-1764	11	-1	Yes
149	5th GGP	Catherine Bowen Gilkey 1712-1784	Catherine [LNU] c1712-c1790	11	-1	Need mtDNA through all females
150	5th GGP	Edward Willis Mercer 1704-1763	Edward Mercer 1704-1763	5	1	Yes
151	5th GGP	Ann Lueretias Coats 1710-1763	Ann [LNU] 1699/1705-c1786/1790	5	-2	Need mtDNA through all females
152	5th GGP	Daniel Brown 1710-1798	Incorrect	39	-2
153	5th GGP	Mary Brown 1717-1777	Incorrect	40	-2
154	5th GGP	Zopher “Elder” Johnson/Johnston* 1700-1804	Incorrect	51	-2
155	5th GGP	Elizabeth Williamson Cooper 1703-1794	Incorrect	49	-2
156	5th GGP	Joseph Benjamin Johnson (6th ggf) (sic) 1709-1795	Incorrect	3	-2
157	5th GGP	Elizabeth Shepard 1709-1786	Incorrect	3	-2
158	5th GGP	John (Boulware) Havis (Rev/war) (sic) 1728-1807	Incorrect	4	-2
159	5th GGP	Susannah Gentile Boullier (Boulware) 1733-1817	Incorrect	3	-2
160	5th GGP	Henry Boulton Jr. 1652-1720	Incorrect	22	-2
161	5th GGP	Elizabeth Bryan 1658-1742	Incorrect, linked in two generations	Duplicate not processing	-2
162	5th GGP	Norton Bryan 1634-1672	Incorrect	2	-2
163	5th GGP	Elizabeth Middlemore 1640-1658	Incorrect	2	-2
164	5th GGP	Guillam Demazure 1685-1706	Guillam Demazares before 1685-after 1705	2	2	Need Y-DNA
165	5th GGP	Marie Demazure 1686-1705	Marie [LNU] before 1686-after 1705	2	1	Need mtDNA through all females
166	5th GGP	John Robert Mann {Minnis} 1711-1772	Incorrect	3	-2
167	5th GGP	Anne Vincent 1711-1747	Incorrect	3	-2
168	5th GGP	Joseph David McCluskey 1693-1756	Incorrect	3	-2
169	5th GGP	Barbara S Rohlflag 1695-1755	Incorrect	3	-2
170	5th GGP	Willis Wilson, Jr. 1710-1794	Incorrect	4	-2
171	5th GGP	Elizabeth Goodrich ?-1789	Incorrect	4	-2
172	5th GGP	Reverend James Matthew Seay 1696-1757	Incorrect	7	-2
173	5th GGP	Elizabeth (James M Seay) Wilson or Lewis 1696-1752	Incorrect	6	-2
174	5th GGP	Ephriam Samuel McDowell 1673-1774	Murtough McDowell before 1700-1752	0	-2	Yes
175	5th GGP	Margaret Elizabeth Irvine 1674-1728	Eleanor [LNU] before 1700-after 1730	1	-2	Need mtDNA through all females
176	5th GGP	Michael Marion Woods 1684-1782	Incorrect	9	-2
177	5th GGP	Mary Catherine Woods 1690-1742	Incorrect	9	-2
178	5th GGP	Joseph Hall 1680-1750	Incorrect	0	-2
179	5th GGP	Sarah Kimball Hall Haley 1686-1752	Incorrect	0	-2
180	5th GGP	Edward Abbott 1702-759	Incorrect	0	-2
181	5th GGP	Dorcas Mehitable Chandler 1704-1748	Incorrect	0	-2
182	5th GGP	James Anderson Clarkston 1717-1816	Incorrect	17	-2
183	5th GGP	Thomasina Elizabeth Middleton 1720-1796	Incorrect	17	-2
184	5th GGP	Harlace Middleton	Incorrect	5	-2
185	5th GGP	Capt. Vallentine Felty Kuke Cook 1730-1797	Incorrect	25	-2
186	5th GGP	Michael Wall 1728-1749	Incorrect	11	-2
187	5th GGP	Rebecca Chapman 1725-1791	Incorrect	11	-2
188	5th GGP	Samuel Scott Muncy 1712-1786	Samuel Muncy 1712-after 1798	50	-1	Yes
189	5th GGP	Mary Daughtery Skidmore 1710-1797	Mary Skidmore c1710-1811	51	-1	Need mtDNA through all females
190	5th GGP	Abraham Woertman Workman 1709-1749	Abraham Workman 1709-1813	26	1	Yes
191	5th GGP	Hannah Annetje (Smith) Workman 1706-1747	Annetie Smith 1714-?	26	1	Need mtDNA through all females
192	5th GGP	Hugh McMahon 1699-1749	Hugh McMahon 1699-1749	17	2	Need Y-DNA
193	5th GGP	Agnas Norton 1699-1747	Agnas Norton after 1700-?	17	2	Need mtDNA through all females
194	5th GGP	Thomas Bowling Speake V 1698-1765	Thomas Speak c1634-1681	11	-2	Yes
195	5th GGP	Jane Barton/Brisco Smoote 1714-1760	Elizabeth Bowling 1641-before 1692	12	-2	No known daughters
196	5th GGP	William Farris 1714-1776	William Faires/Farris before 1728-1776	11	1	Yes
197	5th GGP	Deborah Johnson Faries 1734-1812	Deborah [LNU] 1734-1812	11	1	Need mtDNA through all females
198	5th GGP	Thomas of Borden’s Grant McSpadden 1720-1765	Thomas McSpadden c1721-1785	19	1	Yes
199	5th GGP	Mary Dorothy Edmondson (Edmundson, Edmiston, Edmisten) 1721-1786	Dorothy [possibly Edmiston] 1721-?	28	1	Yes
200	5th GGP	Thomas Alexander McKee, Sr 1693-1769	Incorrect	7	-2
201	5th GGP	Tecumseh Margaret Opessa Pekowi 1695-1780	Incorrect	6	-2
202	5th GGP	Thomas F Nesbit 1707-1783	Incorrect	7	-2
203	5th GGP	Jean McKee 1707-1790	Incorrect	7	-2
	Total				-163

Please note that I will provide a free Y-DNA testing scholarship at FamilyTreeDNA for any male descending through all men from the male ancestor where it’s noted that Y-DNA is needed. Y-DNA is typically the surname line in most western countries.

I will also provide a mitochondrial DNA testing scholarship at FamilyTreeDNA for anyone who descends from the women where it’s noted that mitochondrial DNA is needed. Mitochondrial DNA passes through all females to the current generation, which can be male or female.

If this is you or a family member, please reach out to me.

The Scores

Of the 203 ancestors for which Ancestry provided a Potential Ancestor, they could have amassed a total of 406 points if each one provided an accurate name and accurate birth and death dates within a reasonable margin. If they were completely wrong on every one, they could have earned a negative score of -406.

Ancestry’s ThruLine accuracy score was -163, meaning they were wrong more than right. Zero was the break-even point where there was equally as much accurate information as inaccurate.

In fairness though, the older ancestors are more likely to be wrong than the more recent ones, and there are more older ancestors given that ancestors double in each generation. Once Ancestry provided a wrong ancestor, they continued down that wrong path on up the tree, so once the path was incorrect, it never recovered.

Regardless of why, Ancestry suggested incorrect information, and as we know, many people take that information to heart as gospel. In fact, many people even call these *TrueLines* instead of *ThruLines*.

Ok, how did Ancestry do?

Category	Total	Percent
+2 – Both Name and Date Accurate or Within Range	24	11.82%
+1 – Name and/or Date Partly Accurate	41	20.2%
0 – Uncertain	1	0.49%
-1 – Neither Name nor Date Accurate, but Enough Context to Figure Out With Research	22	10.84%
-2 – Inaccurate, the wrong person	115	56.65%

 Take Aways – Lessons Learned

This leads us to the lessons learned portion.

• Never, ever, take ThruLines or Potential Ancestors at face value. They are hints and nothing more. Ancestry states that “ThruLines uses Ancestry trees to suggest how you may be related to your DNA matches through common ancestors.” (Bolding is mine.)
• Verify everything.
• Never simply copy something from another tree or accept a hint of any kind without a thorough evaluation. No, your ancestor probably did not zigzag back and forth across the country every other year in the 1800s. If you think they did, then you’ll need lots of information to prove that unusual circumstance. Extraordinary circumstances require extraordinary proof.
• Never add extraneous “things” to names like “DNA match” or name someone “Private,” unless, of course, that was actually their name. Extraneous “pieces” in names confuses Ancestry’s search routines too, so you’re hurting your own chances of finding relevant information about your ancestor, not to mention ThruLines for others.
• Naming someone “Private” isn’t useful if they are attached to other non-private people as ancestors, siblings and descendants. Just sayin…
• Once the first incorrect ancestor is suggested, ThruLines continues to go up the incorrect tree.
• In the the older or oldest generations, a small number of DNA matches for a particular ancestor may simply mean that lots of people are beyond the ThruLines match reporting thresholds. Unfortunately, Ancestry does NOT have a function where you can hunt for matches by ancestor.
• In the the older or oldest generations, a small number of DNA matches may also mean it’s either the wrong ancestor, or they have few descendants, or few have tested.
• The number of matches, in either direction, is not directly predictive of the accuracy of the suggested ancestor.
• One of the best ways to validate ancestor accuracy is to match other descendants through multiple children of the ancestor, assuming that the children have been assigned to that ancestor properly. Recall George Middleton Clarkson where the three male children assigned to him do not have the same Y-DNA.
• Another validation technique is to also match descendants of both parents of the ancestor(s) in question, through multiple children.
• Remember that paper trail documentation is an extremely important aspect of genealogy.
• Do not rely on trees without sources, or on trees with sources without verifying that every source is actually referencing this specific person.
• Same name confusion is a very real issue.
• For male ancestors, always check the Y-DNA projects at FamilyTreeDNA to verify that males attached as children have descendants with matching Y-DNA.
• Always test males for their surname line. You never know when you’ll either prove or disprove a long-held belief, or discover that someplace, there has been a biological break in that line.
• Y-DNA matches can provide extremely valuable information on earlier ancestral lines which may lead to breaking through your brick wall.
• Mitochondrial DNA testing and matching of descendants is sometimes the only way of proving maternity or discovering matches to earlier ancestors.
• Both Y-DNA and mitochondrial DNA, via haplogroups, can provide origins information for that one specific line, meaning you don’t have to try to figure out which ancestor contributed some percentage of ethnicity or population-based DNA.
• Everyone can test their mitochondrial DNA, inherited from their direct matrilineal line, and men can test their Y-DNA, which is their surname line.
• Remember that ThruLines can only be as good as the trees upon which it relies.
• Review the source trees for each Potential Ancestor provided, evaluating each source carefully, including notes, images and web links. You just never know where that diamond is hiding.

How Can Ancestry Improve ThruLines, Potential Ancestors and Provide Customers with Better Tools?

To improve ThruLines and/or Potential Ancestors, Ancestry could:

• My #1 request would be to implement a “search by ancestor” feature for DNA matches. This would be especially beneficial for situations where matches are beyond the 5GG threshold, or if someone is testing a hypothesis to see if they match descendants of a particular person.
• Provide a “dismiss” function, or even a function where a customer could provide a reason why they don’t believe a connection or suggestion is accurate. This could travel with that link for other users as well so people can benefit from commentary from and collaboration with others.
• Provide all DNA matches to people who share a specific ancestor, even if one person is beyond the 5 GG level. Currently, if both people are beyond that threshold, the match won’t show for either, so that’s no problem. The hybrid way it works today is both confusing and misleading and the hard cutoff obfuscates matches that have the potential to be extremely useful. Often this is further exacerbated by the 20 cM thresold limit on shared matches.
• Add a feature similar to the now defunct NADs (New Ancestor Discoveries) where Ancestry shows you a group of your matches that descend from common ancestors, but those ancestors are NOT connected to anyone in your tree. However, DO NOT name the tool New Ancestor Discoveries because these people may not be, and often are not, your ancestors. If you’re related to a group of people who all have these people in THEIR tree as ancestors, that alone is a powerful hint. You might be descended from their ancestors, from the spouse of one of their children – something. But it’s information to work with when you have brick walls where Ancestry cannot connect someone as a potential ancestor directly to someone in your tree. Even locations of those brick-wall-breaker possible ancestors would be a clue. In fact, it’s not terribly different than the Potential Ancestors today, except today’s Potential Ancestors are entirely tree based (beyond ThruLines) and dependent upon connecting with someone in your tree. These new Brick-Wall-Breaker Potential Ancestors are (1.) NOT connected to your tree, and (2.) are all a result of DNA matches with people who have these ancestors in their tree.
• If you already map your segment information at DNAPainter, the Brick-Wall-Breaker ancestral lineage connection would be immediately evident if Ancestry provided DNA segment location information. In other words, there are answers and significant hints that could be available to Ancestry’s customers.
• Extend ThruLines for (at least) another two generations. Today ThruLines ends at the point that many people begin running into brick walls about the time the US census began. Using a 25-year generation, the current algorithm gives you 175 years (about 1825 starting with the year 2000), and a 30-year generation gives you 210 years (about 1790). Extending that two additional generations would give testers two more generations, several more Potential Ancestors, and 50-60 more years, approaching or reaching across the US colonial threshold.
• Extending ThruLines and adding that Brick-Wall-Breaker functionality wouldn’t be nearly as important if customers could search by ancestor and download their match with direct ancestor information, similar to the other vendors, but since we can’t, we’re completely reliant on ThruLines and Potential Ancestors for automated connections by ancestor. Downloading your match list including a list of each person’s direct ancestors and matching segments would provide resources for many of these customer needs, without Ancestry having to do significant major development. If nothing else, it could be an interim stepping-stone.

_____________________________________________________________

Follow DNAexplain on Facebook, here or follow me on Twitter, here.

Share the Love!

You’re always welcome to forward articles or links to friends and share on social media.

If you haven’t already subscribed (it’s free,) you can receive an email whenever I publish by clicking the “follow” button on the main blog page, here.

You Can Help Keep This Blog Free

I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.

Thank you so much.

DNA Purchases and Free Uploads

• FamilyTreeDNA – Y, mitochondrial and autosomal DNA testing
• MyHeritage DNA – Autosomal DNA test
• MyHeritage FREE DNA file upload – Upload your DNA file from other vendors free
• AncestryDNA – Autosomal DNA test
• 23andMe Ancestry – Autosomal DNA only, no Health
• 23andMe Ancestry Plus Health

Genealogy Products and Services

• MyHeritage FREE Tree Builder – Genealogy software for your computer
• MyHeritage Subscription with Free Trial
• Legacy Family Tree Webinars – Genealogy and DNA classes, subscription-based, some free
• Legacy Family Tree Software – Genealogy software for your computer
• Newspapers.com – Search newspapers for your ancestors
• NewspaperArchive – Search different newspapers for your ancestors

My Book

• DNA for Native American Genealogy – by Roberta Estes, for those ordering the e-book from anyplace, or paperback within the United States
• DNA for Native American Genealogy – for those ordering the paperback outside the US

Genealogy Books

• Genealogical.com – Lots of wonderful genealogy research books

Genealogy Research

• Legacy Tree Genealogists – Professional genealogy research