Concerns about the confirmed by DNA option

Question

I today changed a medieval relationship which was clearly wrong according to all reasonable sources. When doing so I noted that both the father and mother had been noted as confirmed by DNA. I have some experience with DNA and know that this would have been virtually impossible to confirm both such exact relationships with DNA, and probably would have been published worldwide (and involved digging up the corpses) if it had really happened, so I stuck with using the primary and secondary written sources. It does seem to indicate there might be a problem with this function.

Comments

It's way too easy to set parameters like "Confirmed by DNA." Fortunately, it's also pretty easy to remove that parameter, but when we do that we have to decide between "Confident" and "Uncertain." I'd like to be able to go back to "Unknown" (or more accurately, "Unspecified').

---

Yes, what a worry. I wonder how many profiles this type of thing happens on ?

---

Maybe an unspecified option would also be good for relationships in general, as it seems there is a very big impetus to go for the maximum amount of linking, which is leading to exact-seeming relationships being encouraged. There has to be a better way of preserving speculations without creating misleading information.

---

Would you please enlighten those of us who are new to DNA genealogy? Why is it that you say it is "virtually impossible" to verify a parental relationship dating to the middle ages?  What type or types of DNA tests were cited on the profile in question? My understanding is that Y tests confirming paternal and mitochondrial tests confirming maternal lines are unchanged over generations - as long as you are tracking father-son or mother-daughter relationships through the entire line. C'est nest pas vrai?

---

Michele,

1. Matches which are close enough to indicate a close match in the middle ages are normally simply too common, and "unclose", to do much with them. You can often say that a result looks promising or not promising, but confirming an exact relationship is difficult. Darlene mentioned her father (I was wondering about that!) and how he has a quite unusual type of Y DNA, which is very lucky. (He is also someone I would trust to say whether a type is truly unusual.) But most people are going to find that low level matches are very numerous and might include families with all kinds of names and origins.

2. Let's say you get a reasonably close AND SURPRISINGLY UNUSUAL match between all people expected to be related in the male line, for example all people with a certain surname, or all people in one royal dynasty. OK, what next? Can you therefore say which Smythe is the father of John Smythe of hamlet X? No, just that he looks like he was in the same family, which is not normally going to be surprising.

3. Let's say you get an extremely close match, like 111/111. OK then, you are not talking about a relationship which goes back many centuries then are you?

---

Thank you. I am on the steep, uphill slope of the learning curve. :-)

---

Hi Michele, You are correct that forefather to son yDNA paternal lines and foremother to daughter mtDNA maternal lines extend for millennia.  In fact, they extend much farther back in history than there are records.

Let's take the yDNA test.  If your brother or your father take a yDNA test and have matches with other yDNA tests of the same surname, you can be pretty sure that they share a common yDNA forefather.  The next step is to return to traditional genealogy research to try to find that common forefather, the place where the matching yDNA lines intersect.  The great part is you can ignore all the other lines, at least for now, and concentrate on just those two lines.  In Smith research, this is invaluable!  It is so wonderful to know you are related to this Smith family, but none of those other Smith families.

The complication for medieval connections is there are few written records that can support the yDNA connections.  As an example, we have known for years that William, Simon, Joseph, Benjamin and Edward Smith (born 1618-1633) were genetically related through their yDNA.  In April, we found out that they have the same father. For nearly 400 years, no one knew they were even related, then we got their descendant yDNA, and then we found the marriage record for their parents and the baptism records for all five sons plus five other kids in Stratford-Upon-Avon.  The baptism records support what we know from the yDNA tests.

BUT!! If William Shakespeare hadn't been buried in the chancel of that church, the records probably would never have appeared on the internet and we would never have discovered them and made the yDNA connection.  It was just tremendous luck that Shakespeare was buried in the Holy Trinity Church and my ancestors were married there a week later.

So, for medieval DNA tests, you need matching tests, records that support the matching tests, and a lot of luck finding the written records.

---

Of course the key point here is that the surname Smith has quite a few separate origins, so the different Smith "clans" have yDNA that is different enough for standard tests.

Using yDNA to "confirm" placing within the clan is a whole different matter.

---

Yes RJ but in practice most surnames have many male lines, and most male lines from the same geographical area (with Europe being quite small and homogeneous genetically) will dominated by a small number of lines which did well 1000s of years ago when new technologies expended into Europe and the population exploded. Even with a surname like Lancaster, similar-looking "R1b" lines (the dominant line in most of Western Europe) abound and make it difficult to distinguish families. Mistakes are common (as projects learn as new tests arrive etc) but it is definitely a useful new tool for genealogy. 

(As you say concerning the original subject, distinguishing branches within those identified lines, effectively trying to identify mutations that distinguish exact sets of siblings in the past, is a quantum leap more difficult. Success stories exist, at least when we talk about recent centuries. And I can not recall any such story with mitochondrial DNA.)

Answer 1

This thread has gotten a little confusing, at least for me.  Perhaps we can find an answer on the Confirmed with DNA information page:  http://www.wikitree.com/wiki/Confirmed_with_DNA  

Comments

I find that page very confusing. I think there is way too much room for people to use this feature incorrectly resulting in inaccurate information and conveying to new readers that a relationship has been confirmed as certain when that could be far from the truth.

Answer 2

Which profile was it?  There are some such as Richard III who does have a proven mtdna line, so it's not impossible, but rare as you say.  If it has been tested a source should be quoted in the sources section.

Comments

Hi Veronica, in fact I could not trace back the edit. It is not so easy to trace back past conditions of articles once they are changed apparently.  (I did a lot of edits, and only started to think about it later.) But coming to your remark, no I do not think it is possible that something like the Richard III research can have an effect on making a specific relationship between two people certain, or indeed uncertain. If we are talking about mitochondrial DNA then it is not possible to say whether two matches are close cousins or have a common ancestor centuries ago, so you can have matches from other continents who are as close as known relatives. If we are talking about very high resolution Y DNA tests (not the common 12 or 25 markers) it starts to get a little more sensitive, but basically if you have a very high resolution match you know that two lines are NOT related centuries ago but more recently. Any match which indicates potential relatedness centuries ago is again (like mitochondrial matches) going to be too vague to use to say anything about a specific relationship in a long ago generation. Testing ancient DNA, like you mention, can help of course, but then you need to test several related ancient bodies, not just one, or else you can not triangulate. I guess there will eventually be more such data about royal families, maybe, but I do not think there is much yet.

Answer 3

Here is a link to a similar question I had...

http://www.wikitree.com/g2g/162219/i-have-concerns-with-the-dna-confirmation-help

In Short, I believe the "Confirmed with DNA" should not be a level/mutually exclusive option with "Confident" and "Uncertain".  These refer to the probability that evidence will support a conclusion as to the confidence level of a familial relationship between parent and child.

At what level of probability of DNA is "Confirmed"? auDNA can be ~100%, which I would say is accurate beyond a reasonable doubt, but yDNA and mtDNA can be, In my opinion, compared to auDNA, at best, more likely than not, and leaves significant room for being wrong.

I also believe that wikitree is using this much lower standard, and the result is or will be, influencing people NOT to test, if they believe the test is redundant, when in fact, it is not.

Answer 4

Apparently, if you get a DNA match between two people shown as cousins, you just set the Confirmed flag on every link in the chain.

You believe the line is good because the PMs have signed the Honor Code.

Unsourced profiles?  Well that's the point of DNA isn't it, to prove these undocumented connections.

Comments

Hi Darlene,

The rules you follow are the exact same as those used to "Find" a common ancestor. I believe the rules for "Confirming" a known relationship are different. I am sure you meant to say that persons who share a common grandparent, are 1st cousins.

Just to be sure we are on the same page.  Most Recent Common Ancestor is not a couple, it is a single person. The term came from yDNA and mtDNA comparisons. 

You wrote, "since the original question was dealing with a medieval profile being confirmed with DNA". The point is we are not confirming the MRCA, we are confirming a parent on a profile, one profile at a time, The existing guidelines are designed to somehow confirm a MRCA, not the profiles in the chain.

For some reason, a yDNA comparison between close relatives is not valid. In other words, we shouldn't confirm the fathers on the profiles of the first cousins, even if there is 111/111 match. This is because yDNA comparisons are not intended to "Find" close relatives.

But the guidelines say that if only 1 of those profiles tested against a 10th cousin, then the exact same profile that was rejected with 111 out of 111 markers IS is valid for a 10th cousin who only matches on 100 out of 111 markers.

Not only that, 10th cousin matches, affects 22 profiles, in which we should select "father confirmed by DNA" and include in the biography the EXACT same reason on each of the 22 profiles.

In your case, if you tested 2 male first cousins and the DNA confirms they are full 1st cousins, this confirms they share a common set of grandparents. You won't confirm any profiles, even if they are marked "confident". But the guidelines would have you mark the fathers if they tested against a distant cousin with only 90% of the markers matching.

You comment is not off topic because if I confirmed a 14th cousin by using his yDNA to your brother, we could mark his father as confirmed with DNA, when the same guidelines don't allow it for a confirmed auDNA first cousin.

---

Hello Ken,

You say "For some reason, a yDNA comparison between close relatives is not valid. In other words, we shouldn't confirm the fathers on the profiles of the first cousins, even if there is 111/111 match. This is because yDNA comparisons are not intended to "Find" close relatives."  Perhaps the guidelines need to be revised so they are more understandable.

It makes little sense for close direct paternal line relatives to have a Y-DNA test.  For example it is a waste of money for a father and son to both have a 111 marker test.  Rather the son should Y-DNA test and his  4th, or 5th, etc. direct paternal line cousin should Y-DNA test.  That way if they are a sufficient match then you can indicate each father/son relationship along each tester's direct paternal line back to their most recent (shared) direct paternal line ancestor as confirmed with DNA.

If the son and his 5th cousins don't match as well as they should then you can test closer relatives to determine where the non-paternity event may be.

Sincerely,

---

Hello Ken, I don't understand your statement: " if I confirmed a 14th cousin by using his yDNA to your brother, we could mark his father as confirmed with DNA, when the same guidelines don't allow it for a confirmed auDNA first cousin."  Would you please restate it another way?

A father/son relationship can be confirmed via the Y-DNA and also auDNA of certain relatives.

---

Hi Peter,

Just to be clear, (1) the wikitree guidelines are those based on your personal observations and (2) they are in many cases in direct opposition to FTDNA.

The two responses that come to my mind when asked about using yDNA testing of close relatives to confirm a parent by DNA was not a yes or no, but other reasons. Here is a good example from a previous post that summarizes what I mean. 

Besides it is usually a waste of money to Y-DNA test both fathers and sons because they are expected to match. Otherwise the mother has some explaining to do.

If the purpose of Confirm the father with DNA, then why not just say YES as being within the guidelines. Also, it seems it would be worth every penny spent if we found out the yDNA did not match.

I am not asking "given the choice between yDNA and auDNA, which is better"?, but "When should we select confirm with DNA and when should we not?". 

We can decide for ourselves the financial investment or personal risk on taking these tests.

I believe that knowing when and where mutations occur do have some value and I am willing to promote the purchase of additional kits for closer relatives so that my children and grandchildren may benefit from this, or someone else looking to connect to my tree, can use this to narrow down where in my tree to look. I believe others find this useful as well. You actively discourage this and I don't understand why.

If I have the yDNA results of someone but they have since passed away, which is a scenario that increases every day. Even though it may not be worth the investment to others, It is not a waste of money for many, to also test the son, at a higher level. 

If I already tested my yDNA, but instead of upgrading, I choose to purchase a yDNA test for my father or his brother, I can decided later if I want to upgrade. 

In the examples in this post to make it clearer.

If two first cousins Jack and Bill compare auDNA tests. If the DNA service predicts they are full cousins, nothing is changed on any profile, because their needs to be a third test to triangulate. Correct? 

What is interesting, is that nothing changes if they don't match as 1st cousins either, given the 3 mutually exclusive choices.

To get around this limitation, Bill's yDNA is tested and he matches Tom, who wikitree says are 10th cousins and the DNA comparison indicates they share 90% of their yDNA markers.  

In this case, I should select the "father is confirmed by DNA" on each of the 22 profiles that connect Bill and Tom, and not change anything on the profile of their MRCA. The biography on each of the 22 profiles is updated to state the 90% matching markers between Bill and Tom confirms the father of each of those profiles.  Correct?

This includes the father of Bill, in which the auDNA alone is not enough to confirm. Correct?

We should confirm the father of Bill, even though FTDNA tells us there is "No Relationship" exists between Bill and his 10th cousin Tom. Correct?

Even if the 90% markers can be explained later by finding when and where the mutations occurred, in this particular case, we are on different planets when it comes to calling this "Confirmed" now, without some better explanation of the disagreement between wikitree and FTDNA; and at the same time, not confirming any profiles when the DNA Service predicts a first cousin is a first cousin because some triangulation is necessary.

Confirmed as currently defined, means "more than confident"? Correct

Yet I would have given the current lowest level status of "uncertain" based on FTDNA's probability chart.

The help file should contain links to the FTDNA probability charts and a statement included on that page that FTDNA will in many cases conclude a "No Relationship" when the wikitree 90% rule is applied.

I believe following the existing guidelines will encourage confirmations where they should not be confirmed, and discourage confirmations where they should be confirmed. 

---

Hello Ken, You said "if two first cousins Jack and Bill compare auDNA tests, Nothing on any profile has to be changed.  If the DNA service predicts they are full cousins, nothing is changed on any profile, because their needs to be a third test to triangulate. Correct?"  

There are at least two ways to use auDNA.  Please see:

http://www.wikitree.com/wiki/DNA_Confirmation#Autosomal_Confirmation

Parent child relationships can be labeled Confirmed with DNA when there is a sufficient amount of shared autosomal DNA between two relatives who are third cousins and closer. It is necessary to use Triangulated Groups (link) for confirmation when the testers are greater than third cousins. 

A sufficient amount of auDNA is determined by the total number of cMs of several segments normally shared specific relatives. See Mathematical Average for auDNA Sharing and Are There Any Absolutes in Genetic Genealogy?

Sincerely, Peter

---

Hello Ken,

Regarding different DNA test results for confirming a close family relationship.  My father died before DNA testing was available.  On my direct paternal (Roberts) line I am a 37/37 Y-STR match with my second cousin once removed. I also share the expected amount of autosomal DNA with my father's sister.  I also share the expected amount of autosomal DNA with both of my siblings.  The cM amounts can be seen in GEDmatch and both Y-STR results are in YSearch.  Some of this information is in the citation on my profile for how I know my relationship with my father is confirmed with DNA.  I could add more.

Sincerely, Peter

---

Hi Peter,

I acknowledge the current help infers we might be able to use the predicted relationships.  This seems like a change from the time some people were pushing to require that autosomal data had to be publicly available on Gedmatch so you could determine if there was "when there is a sufficient amount of shared autosomal DNA" and part of the documentation seem to require the shared segment(s) be identified.

This required you to look up this information and each of us could decide what is "Significant". 

A suggested change to the help is to Confirm with DNA if the wikitree relationship is consistent with the DNA service prediction. AncestryDNA does not supply the amount of shared autosomal DNA, regardless of genetic distance.  If the DNA test is consistent, then shouldn't a consistent comparison be enough to confirm with DNA?

In the context of the answer here, Darlene wrote "You need a triangulated match (i.e. three or more people sharing a common segment) that all trace back to a common ancestor in order to confirm by DNA"

She clarified it somewhat, but this seems like the remnants of something previous promoted, and they are not comfortable making the change at least yet. People are still not in many cases, Confirming DNA based on the predicted relationship waiting on a triangulated group. 

---

Hello Ken,  I believe using the DNA service prediction will work for about third cousins and closer if the testers are not from an endogamous group.  If the testers do not provide their GEDmatch ID then it is an honor system. My preference is citations which allow for independent verification like citations to primary source documents enable.

I suggested that WikiTree have different levels of confirmed with DNA.  One level was the DNA testing lab's prediction (with no other evidence).  A higher level included GEDmatch IDs which allowed for independent verification.  The trouble there for some people is that GEDmatch allows registered users to see the e-mail addresses for a member's matches.  

I believe it would be a big mistake to rely on AncestryDNA's New Ancestor Discoveries.   With auDNA, beyond about third cousins I believe we need to rely on triangulated groups.  AncestryDNA's circles and Family Finder's In Common With (ICW) can too easily point to the wrong ancestral couple.

Sincerely, Peter

---

There needs to be more than just Gedmatch, since many people don't use that.  I use DNAGedcom to discover triangulated groups among people that are on FTDNA.  I also use it for Gedmatch (available to tier 1 members).  So you can confirm ICW with both Gedmatch and FTDNA.

As to ancestryDNA, if you've got a large DNA circle there, I believe it can be considered accurate.  One of my circles has 18 members in it.  In looking at the family trees & talking to some of the people, there's no doubt that one is accurate.  I still try and talk people into uploading to Gedmatch, though, so I can see the exact segment(s) we're sharing.  Helps with other matches from FTDNA & 23andMe...

---

Hello Darlene,  I agree that there should be more than GEDmatch.  A problem I see with DNAgedcom is it currently works with in common with (ICW) files (not raw DNA data).  GEDmatch members know (or should know) that others can see their name and contact e-mail.  They joined GEDmatch to make contact with more cousins.  Many of the people in our in common with file did not agree to show who they are and how they match others.

It might be possible for DNAgedcom to get around that problem if they kept a reference database of who (in an ICW file) had agreed to be public.

I believe our time would be much better spent trying to figure out how to get GEDmatch and/or DNAgedcom to allow matches to automatically see their shared ancestry in WikiTree.

Sincerely, Peter

Answer 5

But I'm flummoxed to see how you could confirm both parents of a non-living person, short of digging up the corpse.  Was it a male or female person?

Comments

Digging up a corpse is not necessary.  For example, you take a Y-DNA37 test and your living male "5th" cousin with the Horace surname takes a Y-DNA37 test.  You both match each other on 35 out of 37 markers.  That is a sufficiently close enough match to confirm each of your father/son relationships back to your most recent direct paternal line Horace ancestor.

Sincerely,

---

No Peter, that is not really a confirmation of anything like what the indicator on the wikitree articles implies. What a 37 marker test will show is if 2 people are in a similar Y DNA line. For example more than 50% of English descended male lines will be very close with only 37 markers. And it is quite common to have people with very close matches whose last common ancestor was centuries ago (but it can be much more recent, you don't know). 37/37 is normally within recent centuries, but it still does not tell you where exactly the relationship is. That is really quite difficult and unusual to really do to the point of saying you "confirmed" an exact relationship. People claiming a good triangulation should be asked to publish an explanation of the evidence on the article page in my opinion.

---

I have a question, because I see this all the time....everyone copies everyone else's tree to prove some kind of ancestry.   In that situation, just because 2 people test the same and "say" that they have the same ancestor, is that proof that the ancestor is REALLY proven by DNA?   I think it proves that those two people are related, but, how does it prove their specific common ancestor?

---

I said both parents.  Having "confirmed" that male line to my 5th cousin, how would I confirm any of the dead men's mothers?

The original posting is about a medieval person showing both parents confirmed.  How is that ever possible?

---

Robin your question is a good one to ask, and apparently widely misunderstood. Indeed most DNA results simply give general directions, not exact relationships. For example a reasonably close match between two men with the same surname suggests they are probably from the same male line within a genealogically relevant timeframe. But there are often cases which trick even experienced genetic genealogists. It is basically a question of looking at probabilities. It can be very helpful to genealogy, but it is not a miracle technology which allows us to avoid other types of research. (At least not yet!)

---

RJ, I suppose in theory, as with what is argued for the Richard III case, there might be a claim that a mitochondrial line can be traced all the way to today, and that by luck it is a special unusual one, making it meaningful. In reality... (Anyway this would still only be a test of a whole line, not a specific relationship. Identifying a specific branching point in a family tree, even with Y DNA, requires a lot of luck in terms of mutations, and a lot of triangulation of different branches of descent. It is unfortunately a rare thing to be able to conclude firmly even for recent generations, and even with Y DNA.)

---

But in Richard III's case, they have dug up the corpse.

They couldn't have got his mtDNA from living descendants even if he had any living descendants.

 

---

Right RJH, I believe it was descendants of female relatives of Richard III that were tested. It is a bit of a long shot, but it is defended as believable because (a) the haplotype was apparently unusual and (b) I think they tested several different paper-trail descendants and got the same answer. Very unusual case, at least for now.

---

Hello Andrew, You noted "For example more than 50% of English descended male lines will be very close with only 37 markers."  Please say where you learned that.  I don't believe it is accurate.  There is much more Y haplotype diversity among English males than you note.

My example relates to two males sharing the same (or similar) surname and a direct paternal line ancestry. The 35/37 match confirms that ancestry is accurate.

---

Hello RJ, You are corrrect that a male does not pass any of his mtDNA to his children. So if there had been direct maternal line descendants of Richard III then they would not have his mtDNA.  Richard III's mtDNA, which they uncovered, was from his mother's direct maternal line and they mtDNA tested two living direct maternal line descendants of his mother's direct maternal line.  Richard III's mtDNA matched those two descendants.

---

Peter, sorry the 50% is wrong for 37 markers of course, and of course I was talking about the so called "WAMH" that is much more of a problem below 37 markers, but still a problem at 37 markers. I do not think that changes anything though. In projects I have administered I have had families with the same surname who were still very close at 37 markers but turned out to be different in terms of relatively deep SNPs. Your sentence "The 35/37 match confirms that ancestry is accurate" is simply too vague to mean anything. Which ancestry? As general statement it is not correct. It might be correct in specific cases where other evidence can be drawn upon, as long as we use the word "confirm" in a way which most English speakers will think wrong. A 35/37 match between unrelated male lines is not that uncommon, and even where there is additional reason to think the lines are related, it still says NOTHING about SPECIFIC relationships, and so it can NOT CONFIRM any specific relationship. 

---

Hello Andrew,  My example was 5th direct paternal line male cousins who both have the Horace surname and have a direct paternal line going back to a Horace ancestor.  A 35/37 match confirms that each father/son relationship of those two direct paternal lines back to their shared Horace ancestor is accurate.

A real life example is http://www.wikitree.com/wiki/Bennett-5884 and http://www.wikitree.com/wiki/Bennett-6671 who are 8th cousins once removed from each other on their direct paternal lines:

http://www.wikitree.com/index.php?title=Special:Relationship&action=calculate&person1_name=Bennett-6671&person2_name=Bennett-5884

When you compare them at 37 markers they match each other on 33 out of 37 markers.  At 67 markers they match 63 out of 67 markers:

http://www.ysearch.org/research_start.asp?fail=2&uid=&vallist=NBPZG%2C3E8A8

Prior to Y-DNA testing their genealogy showed that they were 8th cousins once removed.  Now that their haplotypes have been shown to match about 90% or closer, this confirms that their direct paternal line ancestry is accurate back to their most recent Bennett ancestor.

Sincerely, Peter

---

Peter that type of result is not actually a confirmation in the sense most people use the word. Such results are consistent with the paper trail, and can often give us more confidence in a paper trail, especially when looked at with other information. On the other hand two random people with the same surname have a reasonable chance of a close match, and basically the result can only be seen in terms of probabilities. For perspective, on one of the projects I have been administering for many years involves a small Scottish clan. I think there are less than 100 participants who tested through FT DNA. But I have turned off 37 marker match announcements there because there are simply too many, and I still get constant messages about people all over the English speaking world, with a large number of different surnames. This is a well-known type of phenomenum for Scottish clans, which is caused by population expansion over a genetically short period, that is nevertheless outside the timespans we can work with in genealogy.

---

Hello Andrew,

Thanks for your reply.  There is earlier discussion on G2G regarding WikiTree's use of the term "confirmed with DNA."  It is not proof.  Confirmed is a term expert genenetic genealogists have used in relation to autosomal triangulated groups and sufficiently matching Y haplotypes and mtDNA (mitotypes).  Please let me know if you would like me to find their quotes where they use confirmed as it relates to DNA showing that the genealogy is accurate?  Would such quotes change your opinion?

Please also note that many direct paternal lines in Scotland did not have fixed surnames until relatively late.   "The use of  'fixed' (or recognized) Scottish surnames appeared occasionally as early as the 10th or 12th centuries, but didn't begin to be used with any sort of consistency until the 16th century.

Even this, this practice was slow to 'catch on', and it took until the late 18th and early 19th century to spread to the Highlands and northern isles." http://www.scottish-at-heart.com/scottish-surnames.html

Most sincerely, Peter

---

Peter

1.Concerning looking up old discussions your switching to this makes it sound like improving them can not be discussed because discussions are done? I find that way too often on wikitree discussions I have tried to enter so far, and I really do not like that way of talking. I have worked on Wikipedia for many years and this way of talking seems "un-wiki" to me, by which I also mean to imply that it is not based on what has worked in the best wikis that already have more experience. Is this a wiki (with a community that sets its own rules) or some guy's private project to try to get more fake family links than the big commercial sites which we are all helping him with? I really would like to know before I spend any more time on this. But just to repeat: wikitree policies should not be using English words in ways which do not match normal English usage or you will get... situations like the one which started this discussion. I would think that this observation should be interesting to anyone who wants wikitree to be a place with an aim of accuracy. Is wikitree like that or not?

2. As a long time clan project coordinator I am very aware of the way surnames were not fixed until more recently in the highlands. But that is not the only reason for the very dominant haplotypes in Scotland, which are in fact quite old. This was highly discussed for many years given the ethnic make-up of the Anglo Saxon populations who pioneered Y DNA projects. The possibility that this was caused by recent fixing of surnames is something that was checked in many ways by many people. It probably plays a role in some cases. When full sequencing became a realistic possibility this was one of the first things to be double checked. Put simply, the conclusion was already something known to population geneticists, which is that periods of population growth or lack of growth make a massive impact on any attempt to use Y DNA or mitochondrial DNA to predict the age of common ancestors.

Regards

Andrew

---

So can it be definitively stated that in the case of pre-1700 not-dug-up people, the Confirmed flag may only be set on the father of a man or the mother of a woman, never on both parents, because there is no way of "Confirming" cross-gender, in any sense of the word?

---

Well by the standards apparently being used, it seems to be a free-for-all on wikitree, so why not both male and female? The local experts will say that any Y DNA good news means we can paste father-son "confirmations" back for centuries, and any good looking mitochondrial test can allow us to paste mother-child "confirmation" back many generations. (In the case of Richard III, both Y DNA and mitochondrial DNA were tested. From another discussions it seems, though no one confirms it for sure, and unlike on most wikis, on wikitree it is very difficult to trace such things, that Wikitree contributors using the new tool apparently used this to justify calling the relationship between Edward III and John of Gaunt a doubted one.)

Back in the real world confirmations of exact relationships are still quite rare, and apart from cases which are famous enough to be all over the media, they tend to be in very recent generations which are normally going to be set to private on wikitree.

So there seems to be a direct conflict between wikitree policy (or policy interpretation) and reality.

---

RJ, no.  In the case of autosomal DNA and triangulated matches, you can confirm an ancestral couple, as people will match that share that couple as their MRCA (most recent common ancestor), i.e. they will descend from various sons and/or daughters of that couple.  This can be numerous generations back, so you can't say pre-1700, as some people may have their grandparents tested that were born close to 100 years ago, and going back up to 10 generations you can get back to pre-1700 in some situations...

---

Hello Andrew,

From my understanding and perspective WikiTree is not a wiki (with a community that sets its own rules) or as you say some guy's private project to try to get more fake family links than the big commercial sites which we are all helping him with.

My reference to the earlier discussion of confirmed with DNA was not that the topic could not be discussed.  I just wanted you to be aware that there had been prior discussion.  In the earlier discussion of relationship status indicators you can see that I was promoting much more with just "uncertain, confident, and confirmed with DNA."  What we currently have is a start.

My primary concern is that some people are unwilling to include a source citation explaining their selection of confirmed with DNA http://www.wikitree.com/wiki/Confirmed_with_DNA

There are expert genetic genealogists who use the term confirmed when talking about auDNA triangulated groups and using matching Y-STRs and mitotypes as those methods relate to ancestry.  Are you likely to change your opinion if I shared quotes?

WikiTree is using DNA to confirm the accuracy of the known ancestry, not to predict the age of common ancestors.  Are you making reference to "unknown" common ancestors?

Most sincerely, Peter

---

Darlene and Peter, I hope it is clear that I do believe genetic testing can lead to confirmation beyond a reasonable doubt. I think what this discussion seemed to raise doubts about is whether this will be how the function can be used. This is actually a type of evidence which is not easy to get a grip on even for geneticists.

Concerning how to talk about policies Peter, yes pointing to prior discussion can be a good thing, of course. But there is a difference between pointing to a discussion because you just have to do what was said in the past, or pointing to it as a place where there are good explanations. If it was the latter you were trying to do, then I guess that indicates it is hard to explain. And that indicates it will be hard for people to use this function well?

---

A more general comment, maybe for a future discussion.

The wall that common sense seems to keep hitting on G2G discussions always seems to connect a bit to the drive to have lots of links and "cousin bait"? The way this is going it means quantity is being preferred to quality and this is just going to be another cut-and-paste genealogy site like the many others most of us have tried before.

But is this necessary? It seems to be not just about making lots of links, but also about making those links fit in a small number of fields, and that aim is strange. Wikitree uses far fewer fields than gedcoms, which is fine by me, but there also seems to have been a lot of work done to remove functionality compared to normal wiki software. For example articles have no talk pages making it very hard to trace discussions; we are being pushed to designate family connections which are not proven; and we are forced to make strange surnames for medieval people because we are told that "Boolean searching" requires this. The same kinds of answers keep appearing every time I hit something strange.

All these these things are purely artificial decisions, not techical needs. Wikitree has been made this way, and I can not understand why. (Why not create new relationship categories such as "speculative father" in order to reserve the normal relationship functions for clear convincing evidence? And hint to everyone: Google has site searching you can use to search for words not in the main fields. Try searching on google with "site:www.wikitree.com" in the search line. I find it almost impossible to use this website without that. But there are free tools which can install good searching on any website. Normal wikimedia software has normal google-like searching that can look in article bodies.)

---

Hello Andrew,

Thanks for your reply.  Regarding the meaning of confirmed....  At http://www.merriam-webster.com/dictionary/confirm there are two standard definitions:

1. To show that (the parent/child relationship) is correct.

2. To make (the parent/child relationship) more certain.

I believe sufficiently matching Y-STRs, mtDNA (mitotypes), or triangulated groups for autosomal DNA or X chromosome can do both one and two.  I am not referring to proof.

Most sincerely, Peter

---

any good looking mitochondrial test can allow us to paste mother-child "confirmation" back many generations

Not mother-child, only mother-daughter. 

---

I would doubt whether autosomal DNA is good for 10 generations.  Also, the further back you go, the greater the chance of an unknown common ancestor, so the shared segment actually came out of those blank areas behind the brick walls.

---

Mitochondrial DNA goes unmixed, and normally not mutated, from mothers to both daughters and sons.

Concerning autosomal, it is on the way up and still new. There have been surprising results so far. But it is harder to interpret than Y and mitochondrial DNA.

Anyway, as Darlene rightly says, I think, all types of genetic testing for genealogy requiring triangulation of several relatives if anything significant is to be said about a specific relationship in history.

Peter: the above logic applies to both meanings of the word "confirm" which you mention I think, though on the other hand I think it is obvious that there is one meaning most people will understand to be intended.

---

I know a son has his mother's mtDNA.  But he doesn't pass it on, so you can't get at it through his descendants.  You can only get it from the son himself, or what's left of him.

Answer 6

This looks like an opportunity to mention another philosophical concern regarding "confirmed by DNA." I have two different situations in which 3+ fourth cousins (people with the same 3G grandparents) have been found to share a common segment of autosomal DNA (that is, we have a triangulated match). In these instances, we can be reasonably sure that we can trace the DNA segment to a particular couple, but there is no way to tell whether the DNA came from the female or the male ancestor. I could check a "confirmed by DNA" between my 2G grandparent and the couple (i.e., their two parents), but I have no basis for connecting them (or me) to one parent or the other.

I guess I can check the "confirmed" boxes going all the way back to the 2G grandparent, but prior to that I can only mention the relationship in the biography.

Any other ideas on handling this?

Comments

In any case, a match between the 2g-gp's can only tell you they were siblings, it can't tell you who their parents were.

To get anything on that, you'd have to go a step further back and find sibling-matches for the parents.

---

Indeed -- I can confirm by DNA that (using the names from one of these instances) Mary and Martha were sisters, but there is no "confirmed with DNA" option for relationships between siblings.

---

Ellen, You are correct that shared autosomal DNA segments point to a shared ancestral couple.  Unless you find a more distant cousin matching that same segment for those in your triangulated group, who don't know if the shared segment came from the mother or the father and so the relationship with them should not be selected as confirmed with DNA.  The other parent / child relationships back to them should be selected as confimed with DNA with the necessary footnotes.  See the example of a footnote for a confirmation using triangulation at: http://www.wikitree.com/wiki/DNA_Confirmation#Source_Citation_Examples

Sincerely, Peter

Answer 7

My example was 5th direct paternal line male cousins who both have the Horace surname and have a direct paternal line going back to a Horace ancestor.  A 35/37 match confirms that each father/son relationship of those two direct paternal lines back to their shared Horace ancestor is accurate.

Two new words now - "confirms" and "accurate".

Only thing is, me and my 5th cousin made it all up, and all those confirmed ancestors are fictitious.

Let's take a more real example.  Only the names are changed.

It's confidently asserted in "The Pleckett Genealogy" that the Plecketts of Maryland are descendants of Walter Pleckett of the Virginia 3rd Supply Fleet, who (it is claimed) was one of the Plecketts of Plockton.

Could be true, but so could many other things.  There's no evidence.

From statements made when discussing the Uncertain flag, the WikiTree model seems to be that we put those connections in the tree, flagged as Uncertain.  Then we find the descendants and do the tests, and if we get the match we change those Uncertain flags to Confirmed.

Well we're going to get the match, because the Plecketts are all one big clan.  But at the end of it all, we'll be not one jot the wiser about whether those Walter conjectures are true or not.

Obviously there are questions that DNA can answer.  But with male-line cousinships, it seems that Confirmed should only be applied to lines that would otherwise be Confident on paper-trail alone.  If they're Uncertain, yDNA doesn't change that.

The Uncertain help page says

For example, let's say that there is a possibility that Joe Smith's father was John Smith. If there are direct-line male descendants of Joe and direct-line male descendants of John or his direct-line male ancestors, the connection between Joe and John can be reliably confirmed or disproven using Y-chromosome DNA testing.

This seems to be flat wrong and a dangerous misunderstanding.

 

Comments

Well put RJH.

---

Hello RJ,

You noted "Only thing is, me and my 5th cousin made it all up"  Are you saying you and your 5th cousin could make up the genealogy with citations to fictitious evidence and the DNA results to make it all fit the way you want it?  If so, what about your other cousins who are not part of the scheme?  I believe the truth will eventually come to light as more relatives test and check the references to the primary sources.

Sincerely, Peter

---

I saw a family book in which the immigrants (real origin unknown) had been given a fictitious English family to be sons of.

Names and dates allegedly came from a family bible, other stuff from other "family records".  Even extracts from alleged diaries.

All these documents are allegedly in the possession of some unspecified branch of the family.  The writer didn't claim to have them, he only claimed to have seen them.

It's all on WikiTree, in good faith.  Can't remember if the book is cited as the source, but it could be.

All that's needed is to connect the fake family to some real family of the right name that happens to be related to the immigrants' real ancestors, whoever they were.  Then it can all be Confirmed With DNA.

---

Then there's the gerrymandered pedigree in Fitz Randolph Traditions.  It's glorious tosh, but it would pass the DNA test because the male lines do link up further back.

Answer 8

I think that it should be done automatically (ie you can't tick it manually), if a DNA test has been entered.

Comments

The possibility of indicating that a DNA test has been done has been there for a while. What is new is the possibility to note a RELATIONSHIP, for example between John of Gaunt and Edward III (his father as every history book would have it), as confirmed by DNA, or as uncertain. Of course neither of these men were tested.

The two new options are apparently part of a whole package of confusion, because it seems the uncertain option is being used in order to register that there was a DNA test which did not work out as planned.

Answer 9

Another thought.  Suppose I do an autosomal test and discover that I share a segment with somebody I have no known connection to.

Now, the hunt is on for the missing common ancestor.

This sort of thing is liable to lead to all sorts of hypothesizing and handwaving.  But whatever nonsense we come up with, we'll persuade ourselves it's all perfectly plausible, and then say it's confirmed by the shared segment.

Comments

Yes, but such research is not a bad thing. I think your point is that while the debates and discussion go on, we can't be too quickly saying that something has been proven, because autosomal evidence is at the moment complicated to collate and explain. I guess like with Y DNA before it, ways of developing consensus will develop, for example maybe even JOGG or something similar might restart. Maybe websites like wikitree can even help, but with the current formats for showing discussions it would be very difficult here.

---

I think I'm saying that the true reason for the shared segment will often be impossible to find, because the records don't exist, but that won't stop people using traditional methods to come up with something completely wrong.

---

Hello RJ, The hypothesizing and handwaving is why triangulation (for auDNA) is necessary beyond about 3rd cousin relationships.  I would also want to add the need for citations which allow for independent verification.  Otherwise it is too easy to confirm the wrong relationships and it is difficult for others to point out a mistake.

Answer 10

It's interesting to see different views on DNA confirmations.  I am involved in the DNA project at a satellite level - meaning that I am learning and studying and adding my information based on the Guidelines that have been set here on WikiTree.

I will also make this statement.  The DNA guidelines here on WikiTree have been created by people in the project who are not satellite members. We have DNA experts (they make a living by doing DNA work) in the DNA project itself, who have all had input on the standards we use.

My point is this,  the guidelines have been carefully designed to follow the currently accepted standards of DNA research as it relates to Genealogy.

Mags

 

Comments

Well obviously people who work in the industry are the last people we should trust.

There's a propaganda campaign going on.

---

Hi RJ, That last bit is not explained, but I expect that I would disagree, as I think people in the DNA arena are generally knowledgeable and trustworthy about their subject, especially given that there are many different reasons to DNA test and several different types of DNA tests.  I appreciate all the contributors and collaborators on the WikiTree DNA Project.  Kitty Cooper-1 Smith 

Answer 11

Question is, what is it that's being confirmed?

Genetics people are quite glib about what they call paper-trails.  Their attitude tends to be, right, we've got the recorded history, we want to know whether it's telling the truth.  Was the husband really the father?

But genealogists aren't so bothered about that.  They've accepted NPEs as an unavoidable hazard.  Every so often, somebody will say, "of course we don't really know...", and everybody will laugh and carry on regardless.  If you aren't happy with that, might as well give up now.

What worries us is the ambiguity in the records.  I've got this birth cert here, but is it really my ancestor, or is it somebody else?  It doesn't actually say "ancestor of RJ Horace" on it.  My grounds for thinking it's correct may be insufficient.  Many of us have gone off on a wrong trail at one time or another, after persuading ourselves that the wrong person was the right one.

Can DNA help?  Maybe.  But it seems like the connection could be marked as Confirmed when the DNA tests being cited don't really do very much to increase my confidence on the point at issue.

 

 

Comments

Hi RJ,  Re your, "Question is, what is it that's being confirmed?"

The Confirmed with DNA page says, "The "Confirmed with DNA" indicator means that traditional genealogy has been further confirmed with DNA testing."  We have to have the family tree branches in place to support them with the DNA data.

Concerning the balance of your answer/question, I maintain that we never can close the door on additional research and consider a profile "finished".  There can always be errors or misconstructs. We build lots of theories concerning our ancestries, but we have to be aware that a theory is not a proof. New information can change the theory, so we need to be ready to consider any new details.  

---

But most people will not read the help page and they will see "confirmed" and assume all is done.

Others will not read the help page and will select confirmed with DNA once they've submitted a test even if they've done no traditional research at all.

As the original poster pointed out, this feature is being used incorrectly and the results convey inaccurate or at best incomplete info, leading to wikitree perpetuating inaccurate lineages, diminishing its reputation, and frustrating many of us who have worked so hard to IMPROVE the profiles here.