I have no brothers or uncles. Will my DNA results shed any light on my father's line? [closed]

+6 votes
WikiTree profile: Earl Barnes
closed with the note: Question answered 2 years ago
in Genealogy Help by Alta Barnes G2G5 (5.2k points)
closed by Darlene Athey-Hill

3 Answers

+7 votes
Best answer

It could through Autosomal DNA testing. atDNa testing can help on the very near history of your family - just a few generations, but it still could help.

For more information, check out our atDNA help page.


by Mags Gaulden G2G6 Pilot (580k points)
selected by Alta Barnes
Yes, a Y-DNA test on you will show what your father's Y-DNA was. Most of the Barnes line have R1b1a2 Y-DNA.
+5 votes

Hi Alta!  Welcome to our worldwide family WikiTree!

You need to extend these paternal lines of descent and then find a living yDNA Barnes man to do the yDNA test for you. Not easy, but it can be done.  I found a 9th cousin to yDNA test for my Smith line.  http://www.wikitree.com/treewidget/Barnes-775/890  

by Kitty Smith G2G6 Pilot (586k points)
I met a 7th cousin once removed.  We both have Thomas Barnes in common.  His grandmother is his Barnes connection.  Would that work? If I understand this it probably would have to be a male connection to the Barnes line.
For a yDNA test, you would need a males Barnes from a continuous line of male Barnes forefathers.  No women, because women do not have the Y chromosome.

Your distant cousin may know a male Barnes.  You should explain your search and ask him.
Alta -

That's a kit number (you're probably not supposed to quote actual kit numbers for the living, ordinarily, BTW) for autosomal DNA - the whole "direct male descendant" discussion is about YDNA.

A match to somebody believed to be a 7C1R doesn't really prove much, by itself. That distant a relative could be related to you through your mom's side, with some far-back connection you don't even know about. His being male and having the surname doesn't count for much with auDNA. You need more than one hit on the same part of a chromosome, by known relations on different lines, to show that the DNA segment you match on is even one that has been passed down your dad's male line.
I actually really like wikitree's rules for DNA confirmation, it has to be 3rd cousin or closer, or you need additional matches to help triangulate. It's a lot of additional work, but it rules out a lot of false positives.
There is no edit button or I would remove it.  Ancestry said I could download the raw DNA so I thought it would be okay to show it. I don’t really know how to triangulate.  It was explained to me and I want to try and do it but I need to get more information on what I have (the kit numbers) and how to do it.

Thanks...i’ll See if I can remove it.
You are very knowledgeable about DNA so this probably sounds silly or unimportant but a cousin posted a picture of 4 generations of Barnes’ relatives (they are cousins in his tree and my tree.)  One of them could pass for my Dad’s twin.  It’s almost like I don’t need any other verification.  He would be my dad’s 4th or 5th cousin.

Since most or all of the kit numbers I have are probably autosomal, will the triangulation be as valuable or accurate.

Thanks, Alta
I would say triangulation is important if you want to mark your profiles as “confirmed with DNA”.  If you are satisfied with marking profiles as “confident”, then triangulation will be less important to your research.  

For myself, I am happy with “confident” so I don’t concern myself much with triangulation.  That’s my opinion, though others will disagree with me.  

Triangulation is very important if you want to use “confirmed with DNA”.
When you compare a ‘one to many,’ and all those names you see as a (match? to you) how are those matches done? When you get a notice that someone new is a match, how was that discovered? Or derived at?  If you test them with a ‘oneto one,’ it shows a relationship.  Just curious because as say, it’s work to do triangulation. I do want to do that, tho.

I’m happy to hear that we can be satisfied with ‘confidence’ in our matches. I have no close male relatives, so I won’t be able to get too far into it.  I think like everything else, changes might occur that will make all this easier. It’s all very interesting.

Thank you, Alta

There’s no easy way to explain triangulation with an autosomal DNA test.  YDNA is somewhat easier.  I would refer you to two articles I previously wrote here.  https://www.wikitree.com/g2g/229558/how-do-you-use-your-audna-test-on-wikitree  And  https://www.wikitree.com/g2g/544901/do-you-use-dna-painter?show=548446#a548446  

Gedmatch sees your DNA as a string of code. It then is told to look at all the other DNA kits in it's database to see if any of those match yours in any tiny way. Your DNA is measured in centimorgans, abbreviated cM. It gets complicated on that scale, so from my understanding it's more a measurement of information than a physical lenght of DNA. The default gedmatch runs on is matching 7cM or more, because under 7cM can lead to false positives. The more cM match, the closer you are likely related.

https://isogg.org/wiki/Autosomal_DNA_statistics is a very useful tool for figureine out relationships based off of cM amount shared. Gedmatch does have a column in the one to many report called "gen" and that number predicts how many generations back your common ancestor is. It's a little tricky to trust the statistics completely, because most work off of the assumption you inherit 50% of each parent each time, which isn't exactly true. When your parents gametes (egg and sperm cells) developed,they did a thing called recombination or cross over, where it blends thier parents two DNAs together, which is why your dad can give you DNA from both his parents, not just a copy that is 100% one and 0% the other. For instance, my dad and I both match someone else with a 28cM match, that we know is on his mother's side. This must have been a section passed to me intact, without recombination with my grandfather's DNA that halved or shortened it. Gedmatch predicts the same amount of generations back for both my dad to this match, and me to the match, since it's based off how many cMs are shared. I believe that issogg link shows some of the ranges you can get for different relationships.

Since your DNA isn't changing, all new matches should be new kits uploaded that have a string of code that match yours. Gedmatch will highlight these kit numbers in green when you run the one to many test.

Great comment, Allison; thanks. For a somewhat expanded look at a table similar to the one in the ISOGG Wiki, here's one that breaks down theoretical average autosomal DNA sharing out to 7g-grandparents and 8C4R.

For three years now, genetic genealogist Blaine Bettinger has been conducting the Shared cM Project. Here's another useful link to an interactive version of the table where you can enter a total shared amount in centiMorgans and the possible relationships--per the project's data--will be highlighted.

The Shared cM Project is crowd-sourced: people who have autosomal DNA sharing results that match up to known relationships send their data to the project for inclusion. Blaine, scientist that he is, acknowledges the possible accuracy issues of the project data right up front in the full PDF file version of the latest update. Bottom line, each testing company applies different matching thresholds to their results but, and an even greater factor, is that the data are crowd-sourced from user input there's really no way to vet. I'm just noting this because I've seen a recent trend in folks advising to use the Shared cM Project data as gospel, and that the theoretical averages don't matter any longer because we now have some empirical data to work with. Not true.

On the table at that first link above is some limited info from 23andMe about some sharing ranges out to 4th cousins. The theoretical average for 4Cs is 13.3cM; the 23andMe data shows a range of 4.8cM to 34cM; the Shared cM Project shows a range of 0-84cM (95th percentile) and 0-122cM (99th percentile) with an overall average of 36cM...higher than 23andMe's top-end estimate. Inevitably, the crowd-sourced data is going to have far more outliers at the opposite ends of the bell curve...and because the low end will be under-reported (because, well, there are few or no matches for users to report) it's the high end that skews the numbers. This becomes more apparent the more distant the relationship. Much of this can be attributed to pedigree collapse up the tree, with living cousins sharing more total autosomal DNA than expected because somewhere they have fewer biological 2g-grandparents, 3g-grandparents, etc., than the "standard" tree would show.

Net message: thanks for linking to the ISOGG page. The theoretical average autosomal DNA sharing amount is still very relevant, because it helps us frame results that we're seeing. Not long ago I saw a claim of confirmation to a 6th cousin that included a sharing amount 50 times greater than the theoretical amount. The only way that extreme an amount can happen is via pedigree collapse...meaning there was almost a 100% probability that there was more than one ancestral inheritance path that led to the amount of DNA reported, and without thoroughly investigating all those alternate paths, there is no way to accurately assume that cousins A and B got those DNA segments from shared ancestor C. It's the most amazing tool for genealogy since the microfiche  wink  but to use it for relationships beyond 3rd cousins requires time, effort, attention to detail and, most importantly, knowledge.


My initial impression of the big table was very positive - for a start I like the graphics. But then I noticed problems. For a start, it doesn't include "half" relations, except for on mention of half-siblings.

It also says that you share siblings share 3/8 of their DNA, when it's really 50%. The problem is that the cM match calculation isn't a straight % of DNA match, at least not for your very close relations and they have jiggered the shared DNA percentage to match it. When they calculate cMs between you and one of your parents, it comes out near 3400cM because that's a 100% match - they look at both sides of the chromosome and if EITHER matches it counts as a match. Since mom is always on one side or the other, she matches at EVERY location on the chromosome - same for dad. If you did a match against YOURSELF (or a identical twin) you would get the exact same result. You get a 3/4 match with a sibling even though you only have 1/2 the same DNA because there's only a one-in-four chance that NEITHER your DNA from your dad NOR your DNA from your mom matches at a given point on the chromosome. But beyond that, the % actually tracks the cMs. It's a little tricky.

But the calculations they do for about 3C and beyond are meaningless anyway, because they ignore that DNA is inherited in segments. So most of the numbers given are total junk.

Still, I got excited by that blurb in the upper right, which gives info that I had been wondering about. But then I noticed that the info is apparently WRONG. When I look at what it says about 2C, it says they run from 193.8cM to 343.72cM. Half of my own 2C results (AncestryDNA) were below 193 - they go as low as 102! Even Blaine's useful but problematic chart says it goes down to 46cm. I have 2C1R and 3C results that go lower than what they say too.

I've been interested in at what point you will ALWAYS get a match, though, so that part seems interesting and is probably correct. My quandry is that have two matches who I'm pretty sure must be 2C to each other (3C to me), but the Shared Matches list for them say they don't match each other. I know for sure where the one fits on my tree, but not the other (although there things about him that give me a really good idea). It's very puzzling, and he hasn't answered my query.

My brother and I both have the exact same matches up through 2C1R, so I had the impression at one point that that was 100%, but I think I've also seen cases where 2C1Rs didn't, so maybe they're right.

The 89.7% chance of a match for 3C sounds reasonable to me. Both my brother and I match 86% of our 3Cs that I've been able to find are "out there". If I identify some more 3Cs within our matches, my number could go up. But I only know about matches from three places (1) my matches (2) his matches and (3) DNA Circles - so there could also be 3Cs out there that I don't know about, and that would make it go the other way. I don't have enough data of my own to comment on the 4C percentage, but what they have seems reasonable.

I've criticized Blaine's famous table before. Basically, his data analysis is lacking. He includes the endogamous data, which greatly inflates the max numbers (endogamy makes them virtually unlimited). With the distributions being asymmetric, and using the endogamous data, the averages are way too high (they're practically max numbers, for non-endogamous). I don't like the data presentation, with the "halfs" over on the left, and since there are really just "classes" it really doesn't even need to be a 2D table anyway. Oh, and the results are somewhat different for different testing companies, but they're lumped together into one anyway. This is our "Gospel". Sad. Still, it's better than nothing, and I applaud the effort. I'm just saying it's a shame he didn't present it more usefully.

I'd like some real numbers for what's really going on, so that why I've done my own analysis of my own data.

Heya, Frank!  smiley  For the seed info of the "expanded" consanguinity table, including the two methods of looking at sharing distribution, glance back at the link Allison posted to ISOGG: https://isogg.org/wiki/Autosomal_DNA_statistics. In particular, "Method II." I've never had the opportunity to compare full siblings at 23andMe, but "Method II" absolutely holds up on other sites, including GEDmatch. I've worked with over 25 full-sibling and half-sibling matches, and can attest the displayed results are in line with those half-IBD numbers. But, yes: you do not get an average of 37.5% of your DNA from your mother. Leading to...

You're absolutely correct about the vagaries and vicissitudes of centiMorgans. It's far messier even than you describe. We're gettin' way into a rabbit hole of detail that prolly ain't merited by either the age of the original question or its beginner nature, but... 

We've been calculating centiMorgans using the same linear equations that Damodar Kosambi developed decades ago. Putting all the testing company's analysis/reporting differences aside (including that 23andMe adds the xDNA shared amount into the overall, while no other company evens considers it at all), Kosambi's estimation of crossover possibilities based on positioning along a given chromosome has changed very little and is consistent across most all platforms. Thing is, even though the Human Genome Project wrapped up and closed in early 2003, we continue to refine the reference maps we use. And the players don't all use the same version of the map, which can make a significant difference. Kinda like two different evaluations of latitude and longitude. If you and your friend each have a GPS device that does it differently, it probably isn't going to mess things up if you're only going a couple of blocks over. But if the two of you are relying on it to meet for lunch 400 miles away...

Even trickier is that women and men are very different. (I really did just say that with a straight face, honest.) During formation of the human gametes, the rate of DNA recombination is much higher in females than males, as in over 30% higher; some estimates peg it at about 36% higher. The result is that the female genome--while essentially the same size as a male's--looks like it "contains" a boatload more centiMorgans after all the mathematical wizardry. One foundational study on this (Morton, et al. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC52322/) calculated that the resultant female genome is 4,782 centiMorgans long, while the male genome is only 2,809 cM. A little experimentation with the map interpolation tool at Rutgers University bears this out. FTDNA considers each gamete as contributing about 3400cM of material, for a total of about 6800cM; autosomal, no xDNA counted. GEDmatch uses the same assumption of half-IBD. At 23andMe, though--and with the X included--a female genome is considered to be 7439cM, and a male's 7257cM.

So pretty much across the board, we're using estimates, averages, probabilities, population genotyping, and imputation when we look at this stuff. Every cM calculation you see from the testing companies and GEDmatch use a sex-averaged value. Not much more they can do, really. But if, for example, the particular line of ancestral inheritance you're looking at starts with two living males and stays on the patrilineal lines for several generations, you and your cousin are going to display a far smaller amount of autosomal sharing than if the matching cousins are female, and so are all the intervening generations.

Oh, and that bit about the likelihood of sharing detection with cousins came from a study headed up by Brenna Henn, former lead scientist for 23andMe. Like everything DNA, nothing is cast in stone; but I do refer to that study with some frequency when I'm trying to help folks understand the underlying reasons why they, for example, shouldn't be seeing any autosomal sharing with 7th cousins, and that they believe they do is cause for concern and a very deep investigative dive to figure out what's going on...because the MRCA is probably closer than expected, or there's significant pedigree collapse in the line meaning that multiple possibilities exist for the chain of inheritance. You can find that study at PLOS at https://doi.org/10.1371/journal.pone.0034267.

Oops. Sorry, Darlenesad  Just saw that you, rightfully so, closed this question. I went straight to Frank's post from the email link and never caught that. Wish we could actually lock-down a thread when closed.

We'll have to take the conversation elsewhere, Frank. Meet ya in the pub in an hour....

+6 votes
Yes - the Ancestry DNA test could be quite useful for you in identifying possible paternal relatives.

First of all, Ancestry has a database of nearly 1.5 million testers.

Second, it is an autosomal DNA test.  We each receive approximately 50% of our autosomal DNA from our father and approximately 50% from our mother.

Given that your father's family was from Connecticut and England, you may find interesting matches on Ancestry.  Ancestry's database is heavily-concentrated with testers from Colonial America and there are many people with English roots.
by Ray Jones G2G6 Pilot (155k points)
Thank you!
Thank you, I’m hoping to get a better understanding of it, so I can do that.

Related questions

+3 votes
1 answer
164 views asked Nov 23, 2016 in The Tree House by Kelley Harrell G2G6 Mach 1 (16.2k points)
+2 votes
1 answer
+9 votes
2 answers
+8 votes
1 answer
169 views asked Jan 16, 2018 in WikiTree Help by Frank L Barnes III
+2 votes
1 answer
+2 votes
2 answers

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright