Dave, "I was wondering what you were defining as a 'pile-up.'?" I am defining a pileup as an extraordinary number of matches on a particular segment region. These matches are with people I am not related to. Form me, it is chromosome 15 from positions 233,000 to 293,000 where I have several hundred matches. "I am curious if you have crossed checked the reductions against what surnames go missing?" Not yet, in the planning stages after the holidays. "Alternatively, have you counted the number of the pile-up of surnames as well?" Not yet. I have explored enough matches in my pileup region to know that there is no way most, if not all, of these people are related to me. The pileups seem to show up in gedmatch, but it looks like 23andme eliminated my pileup matches in DNA Relatives. David Ann, "You speak of fixing errors -- I assume you're referring to no-calls rather that outright errors (miscalls)?" Yes fixing no-calls. There were only 106 wrong-call between Ancestry and 23. Since I had over 3500 AA, TT, CC and GG that were no-calls in either 23 or ancestry, I would think that a significant number would be opposite homozygote which would break the segment, and reduce the number of matches. David Date: Thu, 17 Dec 2015 15:15:34 -0500 From: Dave Hamm <odoniv@earthlink.net> Subject: Re: [DNA] My Raw Data Files - Comparison 23andme vs, AncestryDNA To: genealogy-dna@rootsweb.com Message-ID: <567317E6.2010300@earthlink.net> Content-Type: text/plain; charset=utf-8; format=flowed Hi David, Oh, I was wondering what you were defining as a 'pile-up.' I did not appear to be having such a problem, and could think of several scenarios of what you could be referring to. I was thinking of your efinition of a 'pile-up' as in the HLA regions. (I identify those with a simple calculation that includes the number of SNPS....) Now I get it. I am now thinking that you are looking at your surname signature(?). Or, at worst, haplotype group signature(?). I get chromosomes with a high density of matching (tiny) segments on chromosomes 1 and 11 when using autosomal sampling from my Y-DNA surname project. I am thinking that I am looking at that as a basic signature for my surname, because of the criteria I use to match each segment triad. Those segments that I do not use in the triad include at least one HLA region on chromosome 6. That HLA region (in my small study) occurs in about 25% of my sampling (after processing). Hence the confusion for me in what you were using for the term 'pile-up.' Apparently, I would gather that a pile-up such as in an HLA region is not what you are referring to. I would think that you may be working on a method that might very well refine your targeted matching segments. I am curious if you have crossed checked the reductions against what surnames go missing? Alternatively, have you counted the number of the pile-up of surnames as well? (I am thinking that losing a random pool of names might be an indication of progress.) Or am I still not understanding the definition (as used here) for a pile-up region? - Dave Hamm Message: 3 Date: Thu, 17 Dec 2015 13:32:05 -0800 From: Ann Turner <dnacousins@gmail.com> Subject: Re: [DNA] My Raw Data Files - Comparison 23andme vs AncestryDNA To: David Schroeder <dschroed991@sbcglobal.net>, DNA Genealogy Mailing List <genealogy-dna@rootsweb.com> Message-ID: <CAA-Ub_AMPi1C0iL97yx4joW7k--BSP99_TszhRM+BjSg1RerRw@mail.gmail.com> Content-Type: text/plain; charset=UTF-8 GEDmatch "tokenizes" the data in an attempt to make it more uniform across platforms, so that should not be an issue for most SNPs. All C's are converted to G's and all A's are converted to T's (or maybe the opposite direction, but you get the idea). That fixes any strand orientation problems, but it introduces a few cases where the alternative alleles are also complementary base pairs in the double helix. The next effect of that is to give you a free ride for those few SNPs where you and your match are really opposite homozygotes. You speak of fixing errors -- I assume you're referring to no-calls rather that outright errors (miscalls). No-calls don't break up segments, just opposite homozygotes (e.g. CC in one party and TT in the other party). GEDmatch doesn't give any credit for the SNP threshold (although 23andMe and FTDNA do). For that reason, I predicted you might have a small increase in the number of matches, and I am puzzled as to why you saw a decrease. Ann Turner On Thu, Dec 17, 2015 at 11:39 AM, David Schroeder via < genealogy-dna@rootsweb.com> wrote:

We're all related to each other, so pileups are still relatives even though it might back a very long time! See the WAAH on CHR 2 as an example Andreas > On 19 Dec 2015, at 02:01, David Schroeder via <genealogy-dna@rootsweb.com> wrote: > > Dave, > > "I was wondering what you were defining as a 'pile-up.'?" I am defining a > pileup as an extraordinary number of matches on a particular segment region. > These matches are with people I am not related to. Form me, it is chromosome > 15 from positions 233,000 to 293,000 where I have several hundred matches. > > "I am curious if you have crossed checked the reductions against what > surnames go missing?" Not yet, in the planning stages after the holidays. > > "Alternatively, have you counted the number of the pile-up of surnames as > well?" Not yet. I have explored enough matches in my pileup region to know > that there is no way most, if not all, of these people are related to me. > The pileups seem to show up in gedmatch, but it looks like 23andme > eliminated my pileup matches in DNA Relatives. > > David > > > Ann, > "You speak of fixing errors -- I assume you're referring to no-calls rather > that outright errors (miscalls)?" Yes fixing no-calls. There were only 106 > wrong-call between Ancestry and 23. Since I had over 3500 AA, TT, CC and GG > that were no-calls in either 23 or ancestry, I would think that a > significant number would be opposite homozygote which would break the > segment, and reduce the number of matches. > > David > > Date: Thu, 17 Dec 2015 15:15:34 -0500 > From: Dave Hamm <odoniv@earthlink.net> > Subject: Re: [DNA] My Raw Data Files - Comparison 23andme vs, > AncestryDNA > To: genealogy-dna@rootsweb.com > Message-ID: <567317E6.2010300@earthlink.net> > Content-Type: text/plain; charset=utf-8; format=flowed > > Hi David, > > Oh, I was wondering what you were defining as a 'pile-up.' > > I did not appear to be having such a problem, and could think of several > scenarios of what you could be referring to. I was thinking of your > efinition of a 'pile-up' as in the HLA regions. (I identify those with a > simple calculation that includes the number of SNPS....) > > Now I get it. > > I am now thinking that you are looking at your surname signature(?). > Or, at worst, haplotype group signature(?). > > I get chromosomes with a high density of matching (tiny) segments on > chromosomes 1 and 11 when using autosomal sampling from my Y-DNA surname > project. > > I am thinking that I am looking at that as a basic signature for my surname, > because of the criteria I use to match each segment triad. > > Those segments that I do not use in the triad include at least one HLA > region on chromosome 6. That HLA region (in my small study) occurs in about > 25% of my sampling (after processing). Hence the confusion for me in what > you were using for the term 'pile-up.' Apparently, I would gather that a > pile-up such as in an HLA region is not what you are referring to. > > I would think that you may be working on a method that might very well > refine your targeted matching segments. > > I am curious if you have crossed checked the reductions against what > surnames go missing? > > Alternatively, have you counted the number of the pile-up of surnames as > well? > > (I am thinking that losing a random pool of names might be an indication of > progress.) > > Or am I still not understanding the definition (as used here) for a pile-up > region? > > - Dave Hamm > > > Message: 3 > Date: Thu, 17 Dec 2015 13:32:05 -0800 > From: Ann Turner <dnacousins@gmail.com> > Subject: Re: [DNA] My Raw Data Files - Comparison 23andme vs > AncestryDNA > To: David Schroeder <dschroed991@sbcglobal.net>, DNA Genealogy > Mailing > List <genealogy-dna@rootsweb.com> > Message-ID: > <CAA-Ub_AMPi1C0iL97yx4joW7k--BSP99_TszhRM+BjSg1RerRw@mail.gmail.com> > Content-Type: text/plain; charset=UTF-8 > > GEDmatch "tokenizes" the data in an attempt to make it more uniform across > platforms, so that should not be an issue for most SNPs. All C's are > converted to G's and all A's are converted to T's (or maybe the opposite > direction, but you get the idea). That fixes any strand orientation > problems, but it introduces a few cases where the alternative alleles are > also complementary base pairs in the double helix. The next effect of that > is to give you a free ride for those few SNPs where you and your match are > really opposite homozygotes. > > You speak of fixing errors -- I assume you're referring to no-calls rather > that outright errors (miscalls). No-calls don't break up segments, just > opposite homozygotes (e.g. CC in one party and TT in the other party). > GEDmatch doesn't give any credit for the SNP threshold (although 23andMe and > FTDNA do). For that reason, I predicted you might have a small increase in > the number of matches, and I am puzzled as to why you saw a decrease. > > Ann Turner > > > On Thu, Dec 17, 2015 at 11:39 AM, David Schroeder via < > genealogy-dna@rootsweb.com> wrote: > > > > ------------------------------- > To unsubscribe from the list, please send an email to GENEALOGY-DNA-request@rootsweb.com with the word 'unsubscribe' without the quotes in the subject and the body of the message