Monday, November 19, 2012

Ancestry.com DNA Research Revealed

imageAncestry.com’s leading DNA scientists participated in the 2012 American Society of Human Genetics (ASHG) annual meeting. Leading DNA experts from across the country gather to present the results of the latest studies. Several scientists from Ancestry presented papers. We can see a bit of the work going on inside Ancestry by reading what they’ve said about their papers:

Pushing the boundaries: Using Haplotypes to infer ancestral origins for recently admixed individuals

This research presents new ways to look at people around the world, and continually pushes our thinking on how we determine ethnicity and population boundaries—specifically in challenging regions like Central Europe—with better data, better algorithms and better analysis.

In other words, Ancestry’s DNA scientists are working with haplotypes to make it possible for AncestryDNA to better determine the ancestral home of people with European ancestry, for example.

Using Y-chromosomes Haplotypes to improve inferred ancestral origins in European populations

In a nutshell, this abstract illustrates how predictions of geographic ethnicity for European populations using autosomal genotypes can be improved by incorporating Y-chromosome information. In fact, using Y-haplogroup distributions to redraw regional boundaries within Europe improved ethnicity predictions by up to 9%.

This description runs counter to what I thought Ancestry was trying to do when it hired these scientists. Beforehand, Ancestry offered a Y-chromosome DNA test—males only—to determine his ancestral home. I thought they hired a bunch of DNA scientists because they were trying to apply autosomal DNA to aid in that determination. This description makes it sound the opposite. It makes it sound like the results of autosomal tests are being refined—again for males only—by going back to the Y-chromosome. This is interesting considering AncestryDNA’s test continues to be available to both men and women and continues to give a full breakdown of genetic ancestry.

Genetic evidence of multiple non-Asian migrations into the new world

An analysis of the Sorenson Molecular Genealogy Foundation (SMGF) Y-chromosome and mtDNA databases resulted in evidence for multiple migrations from the Iberian Peninsula into the New World (Mexico, Central and South America); specifically, two groups were identified—Basque males who share ancestry within the last 2000 years and a Jewish group in Mexico, which fled persecution during the Inquisition.

This study seems to have little application to Ancestry’s business. Perhaps part of the deal when Ancestry acquired Sorenson Molecular Genealogy Foundation (SMGF) DNA samples was to allow some research projects like this one to continue. (One of the samples Ancestry bought from SMGF was my own. I guess the lesson there is to think twice before making a goodwill offering to a non-profit foundation, particularly when the offering is a DNA sample.)

SMGF has some very informative animations teaching more about DNA:

DNA is an exciting frontier in genealogy and it is good to see Ancestry pushing the envelope.

4 comments:

  1. We recently did their test for my dad and were mystified that all of his Germans appeared as Scandinavian--is that because of the Viking invasions? It was hard for me to figure out how far back these results were taking us. And our "Indian ancestor" turned out to be a Southern European percentage, also mysterious since she was born in Nebraska in 1800. I'm glad we tested but it raised more questions than it answered!

    ReplyDelete
  2. DNA testing can not point you to a specific untested ancestor and can not "verify" a paper trail. For example, what if the father of a child some 8 generations back was not a woman's husband but the husband's paternal uncle, even if the husband believed the child was his?

    The autosomal DNA pointers are limited by limited databases of testees to compare with. What if the British Isles database was only from the Orkneys? What if the supposed European database population was based on a few score people who lived in Utah in 2008, whose supposed genealogies were copied from internet trees or (gag) the IGI?

    ReplyDelete
  3. There are several DNA sites that have faulted Ancestry for its conclusions that there are many more Scandinavian ancestors shown on its autosomal test than on other testing sites. This is one comment from DNA-Explained:

    "The problem is that their [Ancestry's] admixture percentages are simply WRONG. Period. Not a “tiny error”, not “needs tweeking,” utterly, entirely wrong. Throw it out and start over wrong. There are no secret Scandinavians hiding in the bushes, or in everyone’s family tree, and the fact that they [Ancestry] are embracing their error and trying to turn a dime by telling people that they DO have a huge amount of mythical Scandinavian blood and they just need to use Ancestry’s tools to search longer and harder is not only infuriating, it’s unethical and self-serving. Several bloggers and others have pointed out that after taking many of these types of tests, Ancestry’s results are the only ones showing large amounts of Scandinavian heritage. So every other company and population geneticist is wrong and Ancestry has made a monumental discovery? Ancestry has been put on notice by many individuals."

    Bottom line is that we shouldn't just trust Ancestry, but should get additional autosomal tests and also follow our paper trails. Ancestry's Scandinavian admixture is inaccurate and the company needs to fix it!!

    ReplyDelete
  4. Ancestry's DNA test results show me as British Isles 97% Uncertain 3%. With names like Smith, Fitzsimmons and Canning as grandparents it's not hard to believe I am up to 75% "British Isles. However, my mother was born a Westphal who's Grandfather was born in Holstein, Germany. Where is the 25% German or "Central European"? The Westphals go back to the 1500's in Geschendorf, Germany. My mothers parents where married for 14 months before she was born so I don't think it was the mailman. My Grandfather was the 5th child of 7 so again probably not the mailman. All of my Gr-Grandfather Westphals family tree contain German names back to the late 1600's or early 1700's so where would there be a "British" connection in this line?

    ReplyDelete