Pedigree collapse deceives us all in our autosomal DNA, au-DNA, matches. I add, one has not really researched their ancestry until they have made themselves to be a cousin to themself at least a couple of ways.
We have a match with a person who shares sufficient chromosomes to makes us 2nd. – 4th cousins but what if we are cousins two ways and both are with common ancestors much farther back.
All these years I get emails from people matching at 4th to distant cousins, more or less, and they are hoping to prove a 17th or 18th century ancestor. I always ask if they have tested other near relatives to identify their matches as paternal or maternal and I usually hear that they have not. We all must work with our nearest matches to identify who they are. We cannot guess with DNA results like we have guessed with trees.
I have more than one ancestor couple who were relatively near cousins, one couple, John Taliaferro Lewis and Eleanor Nelly Earle were 1st cousins. I have in my near autosomal matches, people who appear to be nearer kin than they are because of pedigree collapse.
The way SOME of the companies decide the relationship between two people is how many centimorgans – genes – they share. cM or centimorgan is how genes are valued – like ounces, or pounds, inches or centimeters.
But if I am related to a person through two different families then we might not be 2nd cousins as the numbers of shared centimorgans or genes might show. We might be 5th cousins one way and 7th cousins another way and the total amount of shared genes added up to a number that represents 2nd cousins when in fact the result is deceptive offered as a total.
I love all the bells and whistles of all the DNA testing companies but they all have their limitations and their glitches. This does not make them bad or wrong – but one does have to learn how to read a result.
AncestryDNA has the wonderful tree matching among its DNA participants. They test ONLY au-DNA, they no long give any guessed or presumed haplogroup. And they have the wonderful bells/whistles feature of tree matching. I have called a few times to ask and re-ask how their software decides which ancestor in ones’ trees to match a person to.
They confirmed, time and again, they are not matching the two peoples genes to any given person based on their genetics – they are matching two ancestors in two trees and assuming they must be the common ancestor because the two people share some genes and the two people have the same people in their trees.
We must first hope that the trees are accurate, and that the two people are not related multiple ways. Once one builds up enough matches and the tree circles are created then you are seeing a more accurate picture and a better chance of identifying how two people are related, but there is caution to observe before you would accept a relationship without verifying your relationship a second way. – Especially if you are wanting to identify biological families for those who do not know their ancestry. In other words I am not blowing smoke at people’s hopes – I insist on proving a relationship and then verify it a second way – with another tester.
We know from pedigree collapse that we are related to some families in multiple ways. So we cannot use the total numbers amounts of shared genes, we need to compare those shared genes and see where they are in our genetics.
This is when folks need to ask for help if they are new to this and there are lots of places with people that will help and you not have to pay a genetic genealogist.
My first suggestion is adding another test with another company, this will add the numbers of matches you will possibly have access to, and add more bells and whistles of tools for understanding genetic genealogy results.
This link takes you to an interesting and important story by Alex Shoumatoff
“…The mathematics of descent has fascinated many people. “If we could go back and live again in all of our two hundred and fifty million arithmetical ancestors of the eleventh century,” Henry Adams wrote in 1904 of those with Norman-English blood, “we should find ourselves doing many surprising things, but among the rest we should certainly be ploughing most of the fields of the Contentin and Calvados; going to mass in every parish church in Normand y; rendering military service to every lord, spiritual or temporal, in all this region; and helping to build the Abbey Church at MontSaint- Michel.”…”
“…Pedigree collapse is caused by cousins marrying cousins-:-both intentional mating between close cousins and random mating between distant ones who don’t know that they are related. Close-cousin marriage has happened much more often than is generally supposed. In tribal societies, the exogamic restriction is usually applied not to all one’s blood relatives but only to those in one’s kinship group. In a patrilineal society, for instance, there is nothing to stop one from marrying a matrilineal cousin or uncle; in fact, such a match is often esteemed. The ideal is to marry out, but not too far out. In Japan, which has one of the world’s highest consanguinity rates, arranged marriages between first cousins have been going on for centuries; and surveys made in the nineteen-six ties in southern India found that up to a third of the marriages among the Sudras . of Andhra Pradesh were between first cousins and that the proportion of uncle-niece matings might have been as high as twelve per cent…”
John and Nelly give me many near cousins in matches but those cousins are almost all cousins two ways.
So yes, the autosomal DNA test will in fact pick up some shared genes with cousins who are quite distant. But it is very few of them.
So, let’s say you have paper genealogy between you and a 6th cousin, and you two want to prove this genealogy connection. You both can test and match and you are happy, but I would add another step and make sure the two are related the way we think we are – maybe using a third party for triangulation.
Looking at this from the other side might be best. The two people test and they do not match. Period. They can both even upload to gedmatch and lower the threshold of numbers of total centimorgans and lower the segment length and still not match.
Is it over? I say no. I would add more presumed relatives from both the lines and this is where finding near cousins who would carry the y-DNA of any ancestral lines can efficiently and consistently and accurately be used to prove family lines and cousins – even when they are too far genetically to see in au-DNA results.
I was familiar with other family members’ DNA testing long before I tested myself. I was keen to learn about all the surname DNA projects for my ancestral lines.
The first DNA for genealogy family I learned about was my Lewis family.
See: Colin Powell’s eyes
y-DNA proved that (born 17th century) David Lewis of Hanover through testing male descendants of two of his sons (now four of his sons) matched each other.
Also multiple male descendants of (born 17th century) John Lewis of Gloucester tested their y-DNA and they matched each other.
But the two family groups of males do not match.
What about au-DNA? – but this is where I cringe at the lie we might tell ourselves that au-DNA would in fact, for certain, reach this far back, and prove that you are a descendant of Jean, Julius or Jesus or verify a 10th or 12th generation ancestor.
I have a bridge in Brooklyn, some Snake Oil and links I would like you to click if you can be convinced that your au-DNA result on a pie chart proves you are a Plantagenet.
Now – thing is if you have some iota of European then yes, mathematically you must descend from Charlemagne – but proving it on paper is the trick.
It is such a crock and dishonor to our real ancestors to go off on a 10th generation ancestor that might or might not prove out when in fact we know almost nothing about any ancestor in-between.
One of the great genealogists who we have seen on TV – I must ask her if I can quote her – said something to the effect of the most difficult genealogies to research are where people believe the wrong thing…
From the beginning of the trees on rootsweb and others long before the behemoth Ancestry.com and now other added, those who would listen, learned many if not most of the trees have such flaws – intentional and unintentional – and they have been copied and recopied.
Today two people matching each other with a few centimorgans of genes and having trees that are both made from the same database – you will have a match – but is it identifying the correct family?
Apparently 23andme says they own the idea of looking for matches in a database, based on the amounts of shared centimorgans – so I don’t know if I can even talk about it without begging for permission to tell about finding relatives. They note on their pages to get other relatives, like parents, to test and they will flag your matches based on them being maternal or paternal as other sites can also do.
One does not need software for this just test mother or father or cousins so n so.
I have seen people who match another and insist – oh the match must be here with this 18th century ancestor, or it must be that one with this obscure line from the other place one wishes to prove. But once a person’s matches were sorted by maternal side or paternal side, the match in question often proved to be from the other side of the family where two people thought it was from.
One of my first matches in au-DNA was a lady who was suggested 2nd – 4th cousin + and had in her ancestors surnames, a surname that was prominent in my paternal grandmother Pearl’s neighborhood. I beat the bushes untold hours off and on for a couple of years looking for the link. Several years later a 2nd cousin via only my paternal grandfather Harry tested, and he also matched this lady. So this 3rd person matching both of us proved that the common ancestor was on Harry’s side, not Pearl’s. So testing a third relative to triangulate gave me the answer instantly, where I had been searching on paper futilely for two years.
The same lady used to be in my top 10 closest matches when I first tested, today she is way down on page 5 of my Family Finder matches with 30 matches per page.
I am thrilled to finally see more folks being willing to say the admixture is iffy at best and we all need to see the ethnic results as ‘for entertainment purposes only’ even if it is great entertainment.
…and so we have our ancestries via following haplogroups – mothers and her daughters through time – so once you know your haplogroup you can see your ancestry back in times 1000s of 10s of 1,000s of years ago… and father’s and his sons – and yes i love my admixture also – but DNA is more than a pie chart – it is a passport to Pangaea.
This week another article was shared which also debates some of the same issues.
Adding on June 20, 2018
This article below, by Tina Hesman Saey, has information about admixture and notes how much the results vary – whether it is because the companies have different terms or different – anything – whatever the reason – choose different companies or software to read your admixture.
from this article in Science News they have a 23andme link which I add below to a PDF which includes, that 23andme says it has “50% statistical confidence in the ethnicity results”
Like I always say – I love ALL my admixtures.
great glossary for Gedmatch admixture terms from – and from