On Sunday, May 29, 2016 at 10:02:05 PM UTC-7, taf wrote: > On Sunday, May 29, 2016 at 9:09:32 PM UTC-7, joe...@gmail.com wrote: > > > Your conclusions are flawed and the primary reason is that your references > > are seven to nine years old, which is significant when dealing with new > > technology. > > True. Likewise the original poster cited conclusions based on mtDNA to indict analysis that is based on nuclear genome analysis - apples and oranges. > > > Ancestry.com has built up an enormous database that gives a surprisingly > > good ethnicity estimate that attempts to reach back about 300-500 years. > > Not when my uncle did it. > > > > > I am English, German, French, Italian, Irish, Dutch, etc, etc. and my > > ethnicity (country of origin, really), percentages for each are extremely > > accurate on ancestry.com. > > My uncle's analysis was not good at all, with error bars larger than the average (e.g. 15% Scandinavian, +/- 25% - that is just statistical noise being portrayed as firm data). > > > They have built up a mathematically significant sized database, and > > dedicating increasingly larger amounts of processing power (mainly an > > iterative algorithm) to the problem and making very good progress. > > The problem is not with their statistical or processing power, it is with an oversimplification of what the statistics mean (an oversimplification necessary for mass marketing, but still fundamentally misleading). > > > I accept your conclusions were much more accurate in 2007 or 2009. > > They weren't accurate then either. The original poster has cherry-picked quotes out of context and misapplied them. > > taf REPLY: TAF, God bless you, you never change. What is genetic ancestry testing? This U.S. Library of Medicine article does not cherry-pick or provide quotes out of context and misapply them. https://ghr.nlm.nih.gov/primer/testing/ancestrytesting The American Society of Human Genetics (ASHG) developed a position paper on ancestry testing, with their final testing statement, dated 13 Nov 2008. "Perhaps the most important aspect of reporting confidence in ancestry determinations is to accurately convey the level of uncertainty in the interpretations and to convey the real meaning of that uncertainty." . . . "Population genetic inference is ultimately a statistical exercise, and rarely can definitive conclusions about ancestry be made beyond the assessment of whether putative close relatives are or are not related." . . . http://www.ashg.org/pdf/ASHGAncestryTestingStatement_FINAL.pdf And bringing this forward to 2013, as to putative close relatives. there is "The risk of false inclusion of a relative in parentage testing – an in silico population study", also from U.S. National Library of Medicine, National Institute of Health. "Aim - To investigate the potential of false inclusion of a close genetic relative in paternity testing by using computer generated families." . . . "Conclusions - The results highlight the risk of false inclusion in parentage testing. These data provide a valuable reference when incorporating either a mutation in the father’s DNA type or if a close relative is included as being the father; particularly when there are varying numbers of non-matching loci." Interestingly, "The chance of a false inclusion and exclusion is greater when testing one putative parent and an offspring (a duo scenario) than when there is an additional confirmed parent (a trio scenario)." So, even at this basic level, using "10 000 computer-simulated families over three generations [that] were generated based on genotypes using 15 short tandem repeat loci." . . . "The results highlight the risk of false inclusion in parentage testing." Therefore, TAF, and all other notable senior genealogists: go figure, or calculate, how this negatively plays out with confidence levels, using the more limited DNA testing that actually can be done (using even known paper or compiled professionally documented pedigrees), over time; or pay your own money to read articles presented cheerfully and freely to you, for your easy use, your easy access, and your increased understanding, without posting your own self doubt and personal recriminations. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3692333/ http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3692333/ https://ghr.nlm.nih.gov/primer/testing/ancestrytesting

On Monday, May 30, 2016 at 7:45:32 AM UTC-7, Thomas Milton Tinney, Sr. wrote: > What is genetic ancestry testing? This U.S. Library of Medicine article > does not cherry-pick or provide quotes out of context and misapply them. Just to be clear, this is from the Genetics Home Reference, a set of highly simplified presentation of genetic concepts for a non-scholarly audience. Given that the site you supply doesn't name any sources or provide any quotes, nor does it pretend that a 2007 study of mtDNA in Etruscans and Tuscans is relevant to Ancestry's autosomal testing, this question is just a straw man. > The American Society of Human Genetics (ASHG) developed a position paper > on ancestry testing, with their final testing statement, dated 13 Nov > 2008. As has already been pointed out to you 2008 is ancient history. > And bringing this forward to 2013, as to putative close relatives. > there is "The risk of false inclusion of a relative in parentage testing – > an in silico population study", also from U.S. National Library of > Medicine, National Institute of Health. "Aim - To investigate the > potential of false inclusion of a close genetic relative in paternity > testing by using computer generated families." . . . "Conclusions - The > results highlight the risk of false inclusion in parentage testing. Yet again this compares apples and oranges, but more importantly this study is being misattributed. It is a study by the College of Medicine, National Taiwan University, Taipei, and the Central Police University, Taoyuan, Taiwan, that was published in the Croatian Medical Journal - not exactly a top-flight journal. Indeed, it is the type of journal into which one dumps material just to fluff out one's resume without having to pass rigorous peer review. It has nothing to do with the National Library of Medicine, except that the NLM hosts a database of all published papers from thousands of biomedical journals. To attribute this study to NLM would be like saying that everything in this newsgroup is 'from Google' simply because Google hosts the posts. That being said, the study cited analyzes 15 short tandem repeats and addressed whether in a scenario when a child and mother have tested samples one could unambiguously distinguish the father from a paternal uncle or other close male relative within three generations. Nobody does this analysis on 15 STRs any more, and it is no surprise to me that the best journal they could find to publish this was from Croatia. More importantly, Ancestry does not use STRs at all for their regional ethnicity analyses, and while STRs are used along with single nucleotide polymorphisms for male lineage analyses, these analyses are not trying to do what the journal article is addressing, to definitively conclude which of two brothers might be father of a child. The study is completely irrelevant to the analyses you are trying to condemn. This is made clear from the conclusion. You quote it as saying: "Conclusions - The results highlight the risk of false inclusion in parentage testing." but it continues: "These data provide a valuable reference when incorporating either a mutation in the father’s DNA type or if a close relative is included as being the father" making it clear that either one sentence is being deceptively, intentionally cherry-picked out of the conclusion of an article that is addressing a different question, or else the reader failed to understand the article at all. I will leave it to others to conclude which is the more likely scenario (not that the two are mutually exclusive). > Therefore, TAF, and all other notable senior genealogists: go figure, or > calculate, how this negatively plays out with confidence levels, using the > more limited DNA testing that actually can be done (using even known paper > or compiled professionally documented pedigrees), over time; or pay your > own money to read articles presented cheerfully and freely to you, Still haven't read that article you cited yesterday, have you? taf

On Monday, May 30, 2016 at 10:45:32 AM UTC-4, Thomas Milton Tinney, Sr. wrote: > What is genetic ancestry testing? This U.S. Library of Medicine article does not cherry-pick or provide quotes out of context and misapply them. > https://ghr.nlm.nih.gov/primer/testing/ancestrytesting ULM? No. I'm a little lost on your primary thesis. It seems to be that all "at-home" ancestry tests used for genealogy are fraudulent scams. This is not supported at all by any of your (free) links that you posted whatsoever, or any logical analysis of the evidence. There is a general misunderstanding perhaps about how much ethnic groups are more of a social construct than a biological one, but there is certainly truth in that until recently, there is a strong correlation between how closely two people lived from one another and how closely they were related. The wider disparity, the easier to figure out these clues in the DNA. Can an "At-home" test tell me whether my ancestors were from Dublin or Cork? No. Can they tell Eastern Europeans apart from South-east Asians? Certainly. Can they tell Italians apart from Irish? Also, almost certainly. There is a wider issue of the populations used to calculate these statistical conclusions, and what they are targeting, especially with regard to time period (And the fact they are statistical analysis does not invalidate them as you would imply). It has only been 5 years or less perhaps that general autosomnal testing of any quality has been widely available. Criticizing at-home DNA tests that are based on trying to grab ethnicity out of a Y-DNA or Mitochondrial DNA test may be valid, but it has no bearing on 2016.