Go Back   Family Tree DNA Forums > Universal Lineage Testing (Autosomal DNA) > myOrigins Basics

Reply
 
Thread Tools Display Modes
  #31  
Old 20th February 2018, 08:50 AM
josh w. josh w. is offline
FTDNA Customer
 
Join Date: Mar 2005
Posts: 2,536
Quote:
Originally Posted by josh w. View Post
Maybe they meant that matches go back 500 years. Otherwise, it does not make sense. They classify me as Ashkenazi but that could not be done on East European lines alone. I know that they know that I have a Levantine component (2000 years ago) Display of my Levantine and Italian components is preempted by my designation as Ashkenazi. In other words, if I were not Ashkenazi I could see my Italian and West Middle East components. This is why I need to go to Gedmatch.
If I was not clear, MO could not identify me as Jewish if they did not know my ancient lines. I don't see why adna cannot go back to Viking times. At FF there is a Ydna Viking Project.

Last edited by josh w.; 20th February 2018 at 08:58 AM.
Reply With Quote
  #32  
Old 20th February 2018, 03:42 PM
spruithean spruithean is online now
FTDNA Customer
 
Join Date: Dec 2008
Location: Canada
Posts: 671
Autosomal DNA can't tell you your ethnic origins in the Viking period because it is far beyond 6-10 generations and the chances of sharing centiMorgans with a 6x great-grandparent are slim, and even slimmer (non-existent) in the Viking Age.

We all have pedigree ancestors and genetic ancestors. Read this link.
Reply With Quote
  #33  
Old 20th February 2018, 04:24 PM
John McCoy John McCoy is offline
FTDNA Customer
 
Join Date: Nov 2013
Posts: 691
Autosomal matching depends on detecting long pieces of chromosomes (millions of base pairs) that are identical between or among multiple people. From such long pieces, we can rely on the mechanics of meiosis and recombination to deduce relationships back several generations. If we are very lucky, we can sometimes (rarely) identify relatives as distant as about fifth cousins. Within that range, autosomal matching is found to be remarkably reliable.

Admixture algorithms depend on a completely different set of assumptions: The method assumes the existence of relatively isolated populations that have existed over a long period of time, long enough so that they have accumulated a set of SNP's (mutations, in effect) that are unique to each such population. (We don't actually know that all of the "historical populations" we hope we can measure were ever in fact genetically distinct from each other—that's another of the hopeful assumptions!) The method also attempts to find that set of distinguishing SNP's for each population using statistical techniques applied to DNA samples from living people, who serve as proxies for their long-gone ancestors who belonged to the hypothetical historical populations. The algorithms depend on something like cluster analysis, principal components, or similar multivariate techniques to detect clusters ("reference groups") of people defined by their genetic similarity, and at the same time to detect the sets of SNP's that at once define the "populations" and provide the most reliable way to discriminate among them. But this is far from a simple exercise, for many of the reasons that have been mentioned in this thread. Sample sizes have usually been small, and I strongly suspect they are generally too small to represent adequately the genetic diversity that was present in any real "historical population". It is clear, too, that the "populations" that emerge from the statistical analysis are not distinct enough to produce clear results in many cases. I'm sure the reliability of the algorithms isn't helped, either, by the fact that different vendors are now using very different sets of SNP's for autosomal testing! To the extent that admixture algorithms are being applied to data based on DNA samples that don't contain the same set of SNP's for which an algorithm was developed, we can't expect the results to behave as intended! At best, the admixture algorithms are a work in progress.
Reply With Quote
  #34  
Old 20th February 2018, 06:20 PM
josh w. josh w. is offline
FTDNA Customer
 
Join Date: Mar 2005
Posts: 2,536
Quote:
Originally Posted by John McCoy View Post
Autosomal matching depends on detecting long pieces of chromosomes (millions of base pairs) that are identical between or among multiple people. From such long pieces, we can rely on the mechanics of meiosis and recombination to deduce relationships back several generations. If we are very lucky, we can sometimes (rarely) identify relatives as distant as about fifth cousins. Within that range, autosomal matching is found to be remarkably reliable.

Admixture algorithms depend on a completely different set of assumptions: The method assumes the existence of relatively isolated populations that have existed over a long period of time, long enough so that they have accumulated a set of SNP's (mutations, in effect) that are unique to each such population. (We don't actually know that all of the "historical populations" we hope we can measure were ever in fact genetically distinct from each other—that's another of the hopeful assumptions!) The method also attempts to find that set of distinguishing SNP's for each population using statistical techniques applied to DNA samples from living people, who serve as proxies for their long-gone ancestors who belonged to the hypothetical historical populations. The algorithms depend on something like cluster analysis, principal components, or similar multivariate techniques to detect clusters ("reference groups") of people defined by their genetic similarity, and at the same time to detect the sets of SNP's that at once define the "populations" and provide the most reliable way to discriminate among them. But this is far from a simple exercise, for many of the reasons that have been mentioned in this thread. Sample sizes have usually been small, and I strongly suspect they are generally too small to represent adequately the genetic diversity that was present in any real "historical population". It is clear, too, that the "populations" that emerge from the statistical analysis are not distinct enough to produce clear results in many cases. I'm sure the reliability of the algorithms isn't helped, either, by the fact that different vendors are now using very different sets of SNP's for autosomal testing! To the extent that admixture algorithms are being applied to data based on DNA samples that don't contain the same set of SNP's for which an algorithm was developed, we can't expect the results to behave as intended! At best, the admixture algorithms are a work in progress.
Reply With Quote
  #35  
Old 20th February 2018, 06:23 PM
josh w. josh w. is offline
FTDNA Customer
 
Join Date: Mar 2005
Posts: 2,536
Quote:
Originally Posted by John McCoy View Post
Autosomal matching depends on detecting long pieces of chromosomes (millions of base pairs) that are identical between or among multiple people. From such long pieces, we can rely on the mechanics of meiosis and recombination to deduce relationships back several generations. If we are very lucky, we can sometimes (rarely) identify relatives as distant as about fifth cousins. Within that range, autosomal matching is found to be remarkably reliable.

Admixture algorithms depend on a completely different set of assumptions: The method assumes the existence of relatively isolated populations that have existed over a long period of time, long enough so that they have accumulated a set of SNP's (mutations, in effect) that are unique to each such population. (We don't actually know that all of the "historical populations" we hope we can measure were ever in fact genetically distinct from each other—that's another of the hopeful assumptions!) The method also attempts to find that set of distinguishing SNP's for each population using statistical techniques applied to DNA samples from living people, who serve as proxies for their long-gone ancestors who belonged to the hypothetical historical populations. The algorithms depend on something like cluster analysis, principal components, or similar multivariate techniques to detect clusters ("reference groups") of people defined by their genetic similarity, and at the same time to detect the sets of SNP's that at once define the "populations" and provide the most reliable way to discriminate among them. But this is far from a simple exercise, for many of the reasons that have been mentioned in this thread. Sample sizes have usually been small, and I strongly suspect they are generally too small to represent adequately the genetic diversity that was present in any real "historical population". It is clear, too, that the "populations" that emerge from the statistical analysis are not distinct enough to produce clear results in many cases. I'm sure the reliability of the algorithms isn't helped, either, by the fact that different vendors are now using very different sets of SNP's for autosomal testing! To the extent that admixture algorithms are being applied to data based on DNA samples that don't contain the same set of SNP's for which an algorithm was developed, we can't expect the results to behave as intended! At best, the admixture algorithms are a work in progress.
Agree. My comments about possible Viking related components was in regard to admixture calculations not matches.
Reply With Quote
  #36  
Old 20th February 2018, 06:47 PM
Lani Friend Lani Friend is offline
FTDNA Customer
 
Join Date: Jan 2018
Location: Central Florida
Posts: 34
Quote:
Originally Posted by nygaard View Post
Hello Lani, my mother is as scandinavian as you can be.her mother and father can trace there ancestry back to denmark and norway for more than 400 years.and my dna have me with"only 23% scandinavian dna"
Wow,that's very strange. But it does make me feel better. Have you taken any other tests?

A poster named "Spruithean" Post #18 I think on the second page of this thread recommended two "good reads" in shedding light on this kind of result and I found them excellent. I will post titles here, but you can go to the link on post 18:

Genetic Genealogy using GEDmatch

Understanding Patterns of Inheritance: Where Did My DNA Come From? (And Why It Matters.)
Reply With Quote
  #37  
Old 20th February 2018, 06:52 PM
Lani Friend Lani Friend is offline
FTDNA Customer
 
Join Date: Jan 2018
Location: Central Florida
Posts: 34
Quote:
Originally Posted by josh w. View Post
Maybe they meant that matches go back 500 years. Otherwise, it does not make sense. They classify me as Ashkenazi but that could not be done on East European lines alone. I know that they know that I have a Levantine component (2000 years ago) Display of my Levantine and Italian components is preempted by my designation as Ashkenazi. In other words, if I were not Ashkenazi I could see my Italian and West Middle East components. This is why I need to go to Gedmatch.
Here is what one FT responder wrote to me:
"Anyone who contributed DNA to your genome back farther than five generations will have a greatly diminished overall contribution to your genome that is is most likely too insignificant to be detected or influence your ethnic percentage estimate."

That is why I thought he meant the results don't go back more than several generations. Isn't that what you get out of this statement? Not that it applies just to matches, but that it would apply to percentages also?
Reply With Quote
  #38  
Old 20th February 2018, 07:06 PM
Lani Friend Lani Friend is offline
FTDNA Customer
 
Join Date: Jan 2018
Location: Central Florida
Posts: 34
Quote:
Originally Posted by spruithean View Post
Autosomal DNA can't tell you your ethnic origins in the Viking period because it is far beyond 6-10 generations and the chances of sharing centiMorgans with a 6x great-grandparent are slim, and even slimmer (non-existent) in the Viking Age.

We all have pedigree ancestors and genetic ancestors. Read this link.
This is another fantastic link! In addition to the great links you sent about Genetic Genealogy Using Gedmatch and Patterns of Inheritance. Not exactly light reading, but hugely helpful. Why doesn't FTDNA have these on their webpages?

Question: If this is true, how come the brainiacs at Oxford Univ. were able to use atDNA to construct the phenomenal new "DNA Map of Britain and Ireland" published in journal Nature last year (I think):
http://www.ox.ac.uk/news/2015-03-19-...-british-isles

Those results go back to Celts, Ango-Saxons, Romans, AND Vikings! Why is their autosomal test so different? And doesn't that just prove that atDNA CAN go back to ancient times?
Reply With Quote
  #39  
Old 20th February 2018, 07:14 PM
Lani Friend Lani Friend is offline
FTDNA Customer
 
Join Date: Jan 2018
Location: Central Florida
Posts: 34
Quote:
Originally Posted by josh w. View Post
Agree. My comments about possible Viking related components was in regard to admixture calculations not matches.
Wonderfully detailed info from John McCoy, but my frontal lobe imploded half way through. However, I will study it at length to try to understand. (perils of being a liberal arts person in a genetics forum)
Reply With Quote
  #40  
Old 20th February 2018, 07:30 PM
Lani Friend Lani Friend is offline
FTDNA Customer
 
Join Date: Jan 2018
Location: Central Florida
Posts: 34
Quote:
Originally Posted by josh w. View Post
If I was not clear, MO could not identify me as Jewish if they did not know my ancient lines. I don't see why adna cannot go back to Viking times. At FF there is a Ydna Viking Project.
You were clear. I just didn't understand it correctly. This is all a bit beyond my level of expertise.
Reply With Quote
Reply

Bookmarks


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Segment shared with many Finnish/Norwegian matvhes jbower Family Finder Basics 7 24th February 2017 06:58 PM
Told he was English found out through DNA 1/2 Norwegian LilaJ Adoptees Forum 5 2nd February 2017 06:54 PM
Where did my Scandinavian DNA come from? dcx4610 myOrigins Basics 7 30th August 2016 07:55 PM
Half siblings vs. half aunt/niece Juanita Family Finder Basics 11 11th February 2015 09:19 PM
Norwegian results? henrials Family Finder Advanced Topics 13 17th August 2012 11:34 PM


All times are GMT -5. The time now is 07:50 AM.


Family Tree DNA - World Headquarters

1445 North Loop West, Suite 820
Houston, Texas 77008, USA

Phone: (713) 868-1438 | Fax: (832) 201-7147
Copyright 2001-2010 Genealogy by Genetics, Ltd.
Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.