Emily Yes, this is common. Do the two cousins match each other on the same segment? That is what we are looking for. Jim - Sent from my iPhone - FaceTime! On Nov 20, 2013, at 12:10 AM, "Emily Aulicino" <aulicino@hevanet.com> wrote: > If this fits your situation.... > > I match two of my cousins on the same segment on the same chromosome, but > one is from my mother's line and one from my dad's as they are known cousins > and my parents aren't related. > > If this fits your needs, I can send more detail. > > E >

No...sorry. When compared they don't match as they are on the different pairs of my particular chromo. Sorry. Now I understand. E -----Original Message----- From: autosomal-dna-bounces@rootsweb.com [mailto:autosomal-dna-bounces@rootsweb.com] On Behalf Of Jim Bartlett Sent: Tuesday, November 19, 2013 9:26 PM To: autosomal-dna@rootsweb.com Subject: Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Emily Yes, this is common. Do the two cousins match each other on the same segment? That is what we are looking for. Jim - Sent from my iPhone - FaceTime! On Nov 20, 2013, at 12:10 AM, "Emily Aulicino" <aulicino@hevanet.com> wrote: > If this fits your situation.... > > I match two of my cousins on the same segment on the same chromosome, > but one is from my mother's line and one from my dad's as they are > known cousins and my parents aren't related. > > If this fits your needs, I can send more detail. > > E > ______________________________ For answers to Frequently Asked Questions about mailing lists, please see: http://dgmweb.net/MailingListFAQs.html ------------------------------- To unsubscribe from the list, please send an email to AUTOSOMAL-DNA-request@rootsweb.com with the word 'unsubscribe' without the quotes in the subject and the body of the message

Oops...not on the exact pair of the same chromo, of course! Duh -----Original Message----- From: autosomal-dna-bounces@rootsweb.com [mailto:autosomal-dna-bounces@rootsweb.com] On Behalf Of Emily Aulicino Sent: Tuesday, November 19, 2013 9:11 PM To: autosomal-dna@rootsweb.com Subject: Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing If this fits your situation.... I match two of my cousins on the same segment on the same chromosome, but one is from my mother's line and one from my dad's as they are known cousins and my parents aren't related. If this fits your needs, I can send more detail. E If you do not hear from me in a timely manner, just write again...I was buried in email.  LOL http://writingyourmemories.blogspot.com/ http://www.rootsweb.com/~orgco2/speaker/EmilyAulicino.html http://genealem-geneticgenealogy.blogspot.com/ Northwest Regional Coordinator and Speaker for ISOGG (www.isogg.org) Administrator for thirteen FTDNA DNA Projects -----Original Message----- From: autosomal-dna-bounces@rootsweb.com [mailto:autosomal-dna-bounces@rootsweb.com] On Behalf Of Jim Bartlett Sent: Tuesday, November 19, 2013 9:46 AM To: KATHRYN JOHNSTON; autosomal-dna@rootsweb.com Subject: Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Is there anyone on this list who has experienced a case of three people matching each other on the same segment, and determining that at least two pairs of them have different Common Ancestors? Perhaps we can find some real life examples... Jim - Sent from my iPhone - FaceTime! On Nov 19, 2013, at 12:19 PM, KATHRYN JOHNSTON <kathryn.johnston@sbcglobal.net> wrote: > I don't think there has been a lot of phased studies. It is something > that simply needs to be done. > > > A four way match in an endogamous group could have: > AB > AB > BC > AC > > They all match each other but you should see fully identical matching > for the first two. You would have to look at a chromosome browser that shows the fully identical (double-stranded) segments. This is one more reason to have these kinds of browsers made more readily available by the testing companies. This is the only situation that I can think of that would cause a complete four-way match. > > I think that if everyone matches everyone else and you have a robust > match with all four people with no full matching, and no endogamy or consanguinity, then you are most likely dealing with a single lineage and not the above situation. > > But when dealing with a close-knit heavily drifted community, I would still be on the look-out for these rare birds. I have certainly heard of cases where ALMOST everybody matched everybody else. There could have been threshold issues, mutations, SNP poor regions, errors, IBS etc. instead of IBD endogamy but how would you know if you don't look at the phased raw data and pinned the segment on a specific ancestor? A situation where there are multiple matches, but not sharing with everyone might be: AD, AC, BC, AB, CD, BD. Each of the 6 people matches with four other people and does not match with one. It could fool you into believing that everyone comes from the same MRCA even though there are four ancestors that provided matching segments within the same two position numbers. We all know of groups of ancestors who were highly proliferative. > > It is extremely important to make sure that everyone matches everyone > else as you say when dealing with 4 or more people. But with 3 people, there is no guarantee you have a single ancestral line (even if they do all match each other) without phasing them and/or clear identification of the line. > > People need to study this further in a real life situations. Right now > we are only dealing with a hypothetical problem. > Kathy > > ------ >> Hello Kathy, >> >> What if the ancestry is NOT from an >> endogamous group? What if it is from an endogamous group but is a 4 >> way match? - Where all the matches are matching each other? >> Thanks and sincerely, Peter > ______________________________ For answers to Frequently Asked Questions about mailing lists, please see: http://dgmweb.net/MailingListFAQs.html ------------------------------- To unsubscribe from the list, please send an email to AUTOSOMAL-DNA-request@rootsweb.com with the word 'unsubscribe' without the quotes in the subject and the body of the message ______________________________ For answers to Frequently Asked Questions about mailing lists, please see: http://dgmweb.net/MailingListFAQs.html ------------------------------- To unsubscribe from the list, please send an email to AUTOSOMAL-DNA-request@rootsweb.com with the word 'unsubscribe' without the quotes in the subject and the body of the message

If this fits your situation.... I match two of my cousins on the same segment on the same chromosome, but one is from my mother's line and one from my dad's as they are known cousins and my parents aren't related. If this fits your needs, I can send more detail. E If you do not hear from me in a timely manner, just write again...I was buried in email.  LOL http://writingyourmemories.blogspot.com/ http://www.rootsweb.com/~orgco2/speaker/EmilyAulicino.html http://genealem-geneticgenealogy.blogspot.com/ Northwest Regional Coordinator and Speaker for ISOGG (www.isogg.org) Administrator for thirteen FTDNA DNA Projects -----Original Message----- From: autosomal-dna-bounces@rootsweb.com [mailto:autosomal-dna-bounces@rootsweb.com] On Behalf Of Jim Bartlett Sent: Tuesday, November 19, 2013 9:46 AM To: KATHRYN JOHNSTON; autosomal-dna@rootsweb.com Subject: Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Is there anyone on this list who has experienced a case of three people matching each other on the same segment, and determining that at least two pairs of them have different Common Ancestors? Perhaps we can find some real life examples... Jim - Sent from my iPhone - FaceTime! On Nov 19, 2013, at 12:19 PM, KATHRYN JOHNSTON <kathryn.johnston@sbcglobal.net> wrote: > I don't think there has been a lot of phased studies. It is something that simply needs to be done. > > > A four way match in an endogamous group could have: > AB > AB > BC > AC > > They all match each other but you should see fully identical matching for the first two. You would have to look at a chromosome browser that shows the fully identical (double-stranded) segments. This is one more reason to have these kinds of browsers made more readily available by the testing companies. This is the only situation that I can think of that would cause a complete four-way match. > > I think that if everyone matches everyone else and you have a robust match with all four people with no full matching, and no endogamy or consanguinity, then you are most likely dealing with a single lineage and not the above situation. > > But when dealing with a close-knit heavily drifted community, I would still be on the look-out for these rare birds. I have certainly heard of cases where ALMOST everybody matched everybody else. There could have been threshold issues, mutations, SNP poor regions, errors, IBS etc. instead of IBD endogamy but how would you know if you don't look at the phased raw data and pinned the segment on a specific ancestor? A situation where there are multiple matches, but not sharing with everyone might be: AD, AC, BC, AB, CD, BD. Each of the 6 people matches with four other people and does not match with one. It could fool you into believing that everyone comes from the same MRCA even though there are four ancestors that provided matching segments within the same two position numbers. We all know of groups of ancestors who were highly proliferative. > > It is extremely important to make sure that everyone matches everyone else as you say when dealing with 4 or more people. But with 3 people, there is no guarantee you have a single ancestral line (even if they do all match each other) without phasing them and/or clear identification of the line. > > People need to study this further in a real life situations. Right now we are only dealing with a hypothetical problem. > Kathy > > ------ >> Hello Kathy, >> >> What if the ancestry is NOT from an >> endogamous group? What if it is from an endogamous group but is a 4 >> way match? - Where all the matches are matching each other? >> Thanks and sincerely, Peter > ______________________________ For answers to Frequently Asked Questions about mailing lists, please see: http://dgmweb.net/MailingListFAQs.html ------------------------------- To unsubscribe from the list, please send an email to AUTOSOMAL-DNA-request@rootsweb.com with the word 'unsubscribe' without the quotes in the subject and the body of the message

Is there anyone on this list who has experienced a case of three people matching each other on the same segment, and determining that at least two pairs of them have different Common Ancestors? Perhaps we can find some real life examples... Jim - Sent from my iPhone - FaceTime! On Nov 19, 2013, at 12:19 PM, KATHRYN JOHNSTON <kathryn.johnston@sbcglobal.net> wrote: > I don't think there has been a lot of phased studies. It is something that simply needs to be done. > > > A four way match in an endogamous group could have: > AB > AB > BC > AC > > They all match each other but you should see fully identical matching for the first two. You would have to look at a chromosome browser that shows the fully identical (double-stranded) segments. This is one more reason to have these kinds of browsers made more readily available by the testing companies. This is the only situation that I can think of that would cause a complete four-way match. > > I think that if everyone matches everyone else and you have a robust match with all four people with no full matching, and no endogamy or consanguinity, then you are most likely dealing with a single lineage and not the above situation. > > But when dealing with a close-knit heavily drifted community, I would still be on the look-out for these rare birds. I have certainly heard of cases where ALMOST everybody matched everybody else. There could have been threshold issues, mutations, SNP poor regions, errors, IBS etc. instead of IBD endogamy but how would you know if you don't look at the phased raw data and pinned the segment on a specific ancestor? A situation where there are multiple matches, but not sharing with everyone might be: AD, AC, BC, AB, CD, BD. Each of the 6 people matches with four other people and does not match with one. It could fool you into believing that everyone comes from the same MRCA even though there are four ancestors that provided matching segments within the same two position numbers. We all know of groups of ancestors who were highly proliferative. > > It is extremely important to make sure that everyone matches everyone else as you say when dealing with 4 or more people. But with 3 people, there is no guarantee you have a single ancestral line (even if they do all match each other) without phasing them and/or clear identification of the line. > > People need to study this further in a real life situations. Right now we are only dealing with a hypothetical problem. > Kathy > > ------ >> Hello Kathy, >> >> What if the ancestry is NOT from an >> endogamous group? What if it is from an >> endogamous group but is a 4 way match? - >> Where all the matches are matching each other? >> Thanks and sincerely, Peter >

I don't think there has been a lot of phased studies. It is something that simply needs to be done. A four way match in an endogamous group could have: AB AB BC AC They all match each other but you should see fully identical matching for the first two. You would have to look at a chromosome browser that shows the fully identical (double-stranded) segments. This is one more reason to have these kinds of browsers made more readily available by the testing companies. This is the only situation that I can think of that would cause a complete four-way match. I think that if everyone matches everyone else and you have a robust match with all four people with no full matching, and no endogamy or consanguinity, then you are most likely dealing with a single lineage and not the above situation. But when dealing with a close-knit heavily drifted community, I would still be on the look-out for these rare birds. I have certainly heard of cases where ALMOST everybody matched everybody else. There could have been threshold issues, mutations, SNP poor regions, errors, IBS etc. instead of IBD endogamy but how would you know if you don't look at the phased raw data and pinned the segment on a specific ancestor? A situation where there are multiple matches, but not sharing with everyone might be: AD, AC, BC, AB, CD, BD. Each of the 6 people matches with four other people and does not match with one. It could fool you into believing that everyone comes from the same MRCA even though there are four ancestors that provided matching segments within the same two position numbers. We all know of groups of ancestors who were highly proliferative. It is extremely important to make sure that everyone matches everyone else as you say when dealing with 4 or more people. But with 3 people, there is no guarantee you have a single ancestral line (even if they do all match each other) without phasing them and/or clear identification of the line. People need to study this further in a real life situations. Right now we are only dealing with a hypothetical problem. Kathy ------ > Hello Kathy, > > What if the ancestry is NOT from an > endogamous group?  What if it is from an > endogamous group but is a 4 way match? - > Where all the matches are matching each other? > Thanks and sincerely, Peter

Dear Ann, In situations such as this it is highly probable that the match is IBS. I suspect that in 99% of such cases the match is IBS. However, there is always a possibility that a genotyping error in one of your parents' results or in the match's results is preventing one of your parents from appearing as a match to this person. There is also the possibility that the parent who matches this person doesn't share enough HIRs at least 1 cM in length with this match that total at least 20 cMs. This is another criteria that Family Finder has for a match. Sincerely, Tim Janzen -----Original Message----- From: autosomal-dna-bounces@rootsweb.com [mailto:autosomal-dna-bounces@rootsweb.com] On Behalf Of AGilchrest@aol.com Sent: Monday, November 18, 2013 5:23 PM To: autosomal-dna@rootsweb.com Subject: [AUTOSOMAL-DNA] IBS or IBD Hello, If I have a match that has 27.35 shared cM's with the longest block of 9.01 cM and neither of my parents match my match can I say that this match is IBS? Thank you Ann Gilchrest

Hello, If I have a match that has 27.35 shared cM's with the longest block of 9.01 cM and neither of my parents match my match can I say that this match is IBS? Thank you Ann Gilchrest

Peter I think your statement is correct, and true virtually all of the time. It's an excellent way to set up the draft of your chromosome map. I've now found over 200 Common Ancestors that go with specific segments. This does not include about 30 Common Ancestors I've had to set aside as additional info proves they cannot be correct (they ARE Common Ancestors, they just didn't provide the DNA segment my Match(es) and I share. None of our segments, phased or not, should be locked in until at least one more overlapping segment turns up. And no Common Ancestor should be locked down until at least 1-2 more Matches on that segment can agree on the Common Ancestor. These kind of shifts and confirmations are far more frequent than the very rare chance of an A=B=C=A situation not being Triangulation. At least in my case the score is 30-0. Phasing merely divides the problem into two parts: paternal and maternal. You still have to find the Common Ancestor. The Triangulation process (no phasing required) can divide overlapping shared segments into three groups: two real groups and IBS. All you need is one parent, aunt/uncle, or known cousin to match one of the two real groups to be able to assign them both to the appropriate parent's side. I've been able to assign over 80% of my genome without phasing. This is roughly what Tim has done with phasing. At the end of the day - it's all about determining the Common Ancestors for segments (and confirming them). Jim - Sent from my iPhone - FaceTime! On Nov 18, 2013, at 1:20 PM, peterebay@yahoo.com wrote: > Hello Steven, > > If Alfred matches Betty on a 7cM or greater segment of chromosome 4 and Alfred matches Charles on a 7cM or greater segment of chromosome 4 (which overlaps the Alfred/Betty segment) and Betty matches Charles on this same >=7cM segment, then I believe this almost definitely confirms (or confirms) they all received that segment from a shared ancestor. No phasing is necessary. > > I understand they each have two copies of chromosome 4. > > I would like to hear from anyone as to why the above statement is wrong. > > Thank you and sincerely, Peter > > Peter J. Roberts > > > Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing > Date: Mon, 18 Nov 2013 08:55:29 -0500 > References: <1384781334.49601.YahooMailNeo@web160806.mail.bf1.yahoo.com> > In-Reply-To: <1384781334.49601.YahooMailNeo@web160806.mail.bf1.yahoo.com> > > > Peter: > > All FTDNA tells you is that there is a match on strand A or strand B at a > certain location. Unless you phase the genomes you do not know if they all > match on strand A or strand B or a combination of the two. > > Steven > > > > On Mon, Nov 18, 2013 at 8:28 AM, Peter J. Roberts <peterebay@yahoo.com>wrote: > >> Hello Kethy, >> >> I believe we can skip phasing in situations where three or more people >> share the same segment and ALL match each other on that overlapping/shared >> segment. >> >> Sincerely, Peter >> >> Peter J. Roberts > > ****Please include this message in any replies**** > > > ______________________________ > For answers to Frequently Asked Questions about mailing lists, please see: > http://dgmweb.net/MailingListFAQs.html > > > ------------------------------- > To unsubscribe from the list, please send an email to AUTOSOMAL-DNA-request@rootsweb.com with the word 'unsubscribe' without the quotes in the subject and the body of the message

Hello Kathy, What if the ancestry is NOT from an endogamous group?  What if it is from an endogamous group but is a 4 way match? - Where all the matches are matching each other? Thanks and sincerely, Peter >>>>>>>>>>> From: KATHRYN JOHNSTON <kathryn.johnston@sbcglobal.net> Subject: Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Date: Mon, 18 Nov 2013 12:25:57 -0800 (PST) References: <1384723033.27753.YahooMailNeo@web184301.mail.ne1.yahoo.com><A590520F-F10A-43DA-9F6D-C169A62FBD43@verizon.net> In-Reply-To: <A590520F-F10A-43DA-9F6D-C169A62FBD43@verizon.net> Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Hi Peter, Jim and Steven, I remember that Peter runs a Bahamas project which represents an isolated group of people and Steven has a lot of cousins marrying cousins too so both of you share something in common, endogamy and consanguinity. That is even more reason to consider the possibility that three people can all share with each other and they do not all have to share that match with the same ancestor. There can be more than one ancestral line represented among three people. Yes, phasing is necessary in this situation. And yes, it is very rare but if you are isolated on an island, it may not be that rare. Two copies of chromosome 4 does not mean exact copies. These can come from two different lines. Three people have 6 chromosomes total on chromosome 4. But each person carries two possible haplotypes. They could all carry "A" so match the same ancestor which is usually the case. However, they can carry any combination of A, B and C and not all match exactly the same ancestor but they still COULD match each other on one of the segments that are side by side. The computer program cannot distinguish one segment from another so it just reports it as a match. Alfred may carry haplotype segments A and B Betty may carry haplotype segments B and C Charles may carry haplotype segments A and C The software reads each one as a half-identical-region. The computer does not read which segment you have in common. But this scenario occurs only when each person has some combination of A, B and C. If a new person comes along with D plus A, B or C, then he will not match somebody but will match two out of the three people above. Regards, Kathy

Hello Steven, If Alfred matches Betty on a 7cM or greater segment of chromosome 4 and Alfred matches Charles on a 7cM or greater segment of chromosome 4 (which overlaps the Alfred/Betty segment) and Betty matches Charles on this same >=7cM segment, then I believe this almost definitely confirms (or confirms) they all received that segment from a shared ancestor. No phasing is necessary. I understand they each have two copies of chromosome 4. I would like to hear from anyone as to why the above statement is wrong. Thank you and sincerely, Peter Peter J. Roberts >>>>>>>>>>>>> Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Date: Mon, 18 Nov 2013 08:55:29 -0500 References: <1384781334.49601.YahooMailNeo@web160806.mail.bf1.yahoo.com> In-Reply-To: <1384781334.49601.YahooMailNeo@web160806.mail.bf1.yahoo.com> Peter: All FTDNA tells you is that there is a match on strand A or strand B at a certain location. Unless you phase the genomes you do not know if they all match on strand A or strand B or a combination of the two. Steven On Mon, Nov 18, 2013 at 8:28 AM, Peter J. Roberts <peterebay@yahoo.com>wrote: > Hello Kethy, > > I believe we can skip phasing in situations where three or more people > share the same segment and ALL match each other on that overlapping/shared > segment. > > Sincerely, Peter > > Peter J. Roberts > ****Please include this message in any replies****

Re: [AUTOSOMAL-DNA] Finding shared atDNA ancestry without phasing Hi Peter, Jim and Steven, I remember that Peter runs a Bahamas project which represents an isolated group of people and Steven has a lot of cousins marrying cousins too so both of you share something in common, endogamy and consanguinity. That is even more reason to consider the possibility that three people can all share with each other and they do not all have to share that match with the same ancestor. There can be more than one ancestral line represented among three people. Yes, phasing is necessary in this situation. And yes, it is very rare but if you are isolated on an island, it may not be that rare. Two copies of chromosome 4 does not mean exact copies. These can come from two different lines. Three people have 6 chromosomes total on chromosome 4. But each person carries two possible haplotypes. They could all carry "A" so match the same ancestor which is usually the case. However, they can carry any combination of A, B and C and not all match exactly the same ancestor but they still COULD match each other on one of the segments that are side by side. The computer program cannot distinguish one segment from another so it just reports it as a match. Alfred may carry haplotype segments A and B Betty may carry haplotype segments B and C Charles may carry haplotype segments A and C The software reads each one as a half-identical-region. The computer does not read which segment you have in common. But this scenario occurs only when each person has some combination of A, B and C. If a new person comes along with D plus A, B or C, then he will not match somebody but will match two out of the three people above. Regards, Kathy --------------- 
> From: peterebay@yahoo.com 
> Subject: Re: [AUTOSOMAL-DNA] Finding shared > atDNA ancestry without phasing 
> Date: Mon, 18 Nov 2013 13:20:34 -0500 > > Hello Steven, > > If Alfred matches Betty on a 7cM or > greater segment of chromosome 4 and Alfred > matches Charles on a 7cM or greater > segment of chromosome 4 (which overlaps > the Alfred/Betty segment) and Betty > matches Charles on this same >=7cM > segment, then I believe this almost > definitely confirms (or confirms) they all > received that segment from a shared > ancestor. No phasing is necessary. > > I understand they each have two copies of > chromosome 4. > > I would like to hear from anyone as to why > the above statement is wrong. > > Thank you and sincerely, Peter > > Peter J. Roberts

Peter: All FTDNA tells you is that there is a match on strand A or strand B at a certain location. Unless you phase the genomes you do not know if they all match on strand A or strand B or a combination of the two. Steven On Mon, Nov 18, 2013 at 8:28 AM, Peter J. Roberts <peterebay@yahoo.com>wrote: > Hello Kethy, > > I believe we can skip phasing in situations where three or more people > share the same segment and ALL match each other on that overlapping/shared > segment. > > Sincerely, Peter > > Peter J. Roberts > > > ______________________________ > For answers to Frequently Asked Questions about mailing lists, please see: > http://dgmweb.net/MailingListFAQs.html > > > ------------------------------- > To unsubscribe from the list, please send an email to > AUTOSOMAL-DNA-request@rootsweb.com with the word 'unsubscribe' without > the quotes in the subject and the body of the message > -- Steven C. Perkins SCPerkins@gmail.com http://stevencperkins.com/ Indigenous Peoples' Rights http://intelligent-internet.info/law/ipr2.html Indigenous & Ethnic Minority Legal News http://iemlnews.blogspot.com/ Online Journal of Genetics and Genealogy http://jgg-online.blogspot.com/ S.C. Perkins' Genealogy Page http://stevencperkins.com/genealogy.html S.C. Perkins' Genealogy Blog http://scpgen.blogspot.com/

I agree with Kathy that the scenario she described is very rare. This would involve three 7cM segments that were exactly alike over the whole area from 3 separate ancestors. It is possible. It is possible to have 20 and 30cM shared segments that are IBS. But very unlikely. With the randomness of DNA, unusual events occur (a few people exist with 3 sex chromosomes: XXY - and other combinations) For general spreadsheet analysis and chromosome mapping, I think it's safe to consider all of your matches on a particular segment to come from one of two of your ancestors (one paternal, one maternal), and any who match each other are Triangulated, and come from the same ancestor. I'm now curious as to why the original question came up: what is > >> the likelihood they don't all descend >> from the same ancestor... It seems to me that if a fit cannot be found, there is a much more likelihood that the answer is behind a genealogy brick wall (or mistake), than a rare genetic event. IMO. But with DNA strange things will happen... Jim - Sent from my iPhone - FaceTime! On Nov 17, 2013, at 4:17 PM, KATHRYN JOHNSTON <kathryn.johnston@sbcglobal.net> wrote: > To answer Peter's first question below. You asked about finding shared segments without phasing which means you don't know which one out of the two segments you are sharing on. So that means three people can share with each other without all sharing on the exact same segment even though the segments line up in exactly the same places. Each person can match one person on the segment coming from his/her mother and also match the other person on the segment coming his/her father. I call it a three-way but not exact match. > > Example, John, Kate and Mary all match each other at approximately 7 cM on a chromosome browser, in the same spot and the segments are IBD. John and Kate match the German ancestor here. John and Mary match the Irish ancestor here. Kate and Mary match the English ancestor here. There are three ancestral lines. Each person only matches two out of the three. Mary may not have any German ancestry and does not descend from that ancestral couple yet she still matches the other two people. > > It is like holding hands in a circle with two other people. There are three different matching possibilities and you don't match where the two other people are holding hands. But you all connect with each other. I think this is very rare but can certainly happen in communities where the ancestors were all living in nearby circles (pun intended) but not all related to everybody else through exactly the same ancestors. > > > Kathy J. > > -- >> On 11/15/13, Peter J. Roberts<peterebay@yahoo.com> > wrote: >> >> If three or more people share a 7cM segment >> and they all match each other on >> that segment then what is >> the likelihood they don't all descend >> from the same ancestor or >> ancestral couple? >> What would be other exceptions >> besides double cousins? >> Thanks and sincerely, >> Peter >> Peter J. Roberts

Hello Kethy, I believe we can skip phasing in situations where three or more people share the same segment and ALL match each other on that overlapping/shared segment. Sincerely, Peter Peter J. Roberts

Denise, I went through the same feelings of unease when using the 23andMe tools vs the FTDNA analysis tools. What changed that all about was Rob Warthen's DNAgedcom.com utility. Since I can now download all of my CB matches for the accounts I manage on FTDNA, and I can also download the data files from 23andMe for those accounts as well into spreadsheets, * * *23andMe is the clear winner (IMHO) for ICW analysis* -- but with caveats. The most useful file in the 23andMe set is the AF file, which shows at the chromosome level, those with whom you share genomes and public matches by name (all others are Anonymous###, where ### is a number up to 0999 in my case) and those whom have filled out the birthplaces of the grandparents in the Countries of Ancestry Survey. Those are the caveats. That is, someone has to either have listed their profiles as public matches or shared genomes, but all have to fill out the CofA survey. Since I have a multiple profile 23andMe account I get a very large csv AF file -- usually North of 25 mB to download from DNAgedcom.com -- with all of the chromosome level matches for all of my profiles at one swell foop...er...fell swoop. This results in a very large spreadsheet when opened in Excel (of late 177,829 rows by 15 columns). {Note: my computer's ability to handle these large data arrays improved remarkably, when I switched to a 64-bit machine (running Windows 8) with 12GB of RAM and most importantly installed a 64-bit version of Excel. Most of my Excel demons went away with that change. I could even run Tim Janzen's manual phasing routine for three raw data arrays with no problems whatsoever after installing the 64-bit Excel.} The first thing I do is to open the csv file from 23andMe on DNAgedcom.com and then "Save as" an .xlsx file immediately. Otherwise Excel will complain of data loss, if you start to customize the file while in .csv form and then save it later. Next, I generally either delete or hide eight of the columns (C to J) in the 23andMe AF spreadsheet namely: MaternalGrandmotherBirthCountry, MaternalGrandfatherBirthCountry, PaternalGrandmotherBirthCountry , PaternalGrandfatherBirthCountry , MaternalGrandmotherDeclaredAshkenazi, MaternalGrandfatherDeclaredAshkenazi, PaternalGrandmotherDeclaredAshkenazi , and PaternalGrandfatherDeclaredAshkenazi. Sometimes I also delete the SegmentLengthInMegaBasePairs column (column N) as well as it is really not needed. Then I quickly shorten the names of the remaining columns from Chromosome, SegmentStartInMegaBasePairs, SegmentEndInMegaBasePairs, and SegmentLengthInCentiMorgans to Chr, Start, End, and cM. This clean up sounds like a lot, but it actually takes less than 30 seconds to do all of it. In this example, the above dimensions of the abridged worksheet are now 177,829 rows by 6 columns and contain the Ancestry Finder profiles of all seven people in my account along with all of their individual matches and everyone with whom we are sharing genomes! Lastly I highlight all the remaining columns and then set Filters on all of them. The Filter function is a most wonderful tool! It allows one to cull and highlight data in an amazing variety of ways. I usually set the cM filter to >7 cm, and set the Matchname filter to "Does not contain" Anony*, where the * is a wildcard. This removes a thousand Anonymous entries and a great deal of the IBS entries, which in my example above amounts to 65,588 entries > 7 cM. Setting the cM filter to >11 is even more effective in getting rid of superfluous entries and reduces the total to some 53,705. This is a far cry from the original 177,829 rows. Sorting the cM column from largest to smallest then allows one to focus on the most significant matches. Increasingly, I am finding that I rarely get anything of value from pursuing smaller matching segments. These small block matches are to more distant cousins, who when contacted usually do not know their pedigrees very far back. And even if one or two descent lines are known, there remain many gaps in the other lines. Similar efforts are required to pursue these small matches as compared to the larger, more significant blocks. While there is give and take in terms of data sharing, I find that some people want you to do their genealogy for them as opposed to sharing data in common pursuit of pedigrees. My advice is to pick the low hanging fruit first and then if there is a multiple match on a given segment, pursue the smaller block matches last. When I identify a block on a chromosome that I want to work on, I then limit the Chr column filter to that specific chromosome only.** _ __As long as you are sharing genomes or the match is public listed on 23andMe, you can see all of their matches as well._ The net result is that you get a true, chromosome level, block specific match display. A true and very useful "In Common With." In this regard, the 23andMe AF data + DNAgedcom.com beats FTDNA hands down even with their recent "improvements." (The shame of it is that it would be so easy for FTDNA to trump this with a few simple and long sought changes: 1.) A true, specific chromosome level ICW, and 2.) a way to download all CB data of that ICW match to an Excel spreadsheet.) ****** It is also possible to put together in one uniform worksheet, the results of an FTDNA CB download and the 23andMe AF data with common matches shown in one worksheet. I do that quite frequently. The caveat there is that you need at least two people who are on both platforms to make the ICW really work and then because one cannot see if there is an A=B=C match in the FTDNA data, that is the only lame thing in those displays. (FTDNA are you listening? Rebeka? Anyone?) I have had some really notable successes in doing this. Recently I triangulated a Chr 4 block to one set of my great great great grandparents, with some 14 people identified either on 23andMe and/or FTDNA who match! There were 11 original but since doing the initial correlation, three more have shown up from the 23andMe AF matches. Three of this set of 14 triangulate to 4C and one to 4C2R, with the balance undetermined, but definitely connected. There is one known NPE and one adoption in the group of 14, BTW. This is an on-going project and I anticipate another report in coming months, which is only waiting for some of the slow or no responders to get back to me. Based on more recent work, I have hints and a hunch, that I may be able to identify to which of the ancestral couple this block of common DNA on Chr 4 is owing! And a further lead on yet another block on Chr 18 which may be owing to the other partner of the ancestral couple. If I can sanitize my initial analysis report of personal names, I would be willing to share a copy of that as a Word File on request. Walter ** I have a simple filter routine that I also perform on the Start and End settings of an individual block, which shows only those people who match both me and if they match one another or not. I am stopping short here of describing that procedure owing to the length of this post. However, I will explain in detail how that works on request (on or off list as you prefer.) On 11/16/2013 8:28 PM, Denise Sproed wrote: > On Family Tree DNA one can 'triangulate' for 'in common matches' (not > necessarily in common on the same chromosome)... > > Does 23andme have a similar tool? Or do folks all download the data from > www.dnagedcom.com so they can manually review which people match on a > specific chromosome? > > I am aware Family Inheritance Advanced lets you compare 3 people to another > but that doesn't tell you who ELSE in the list might share that same > chromosome location... > > Denise > >

To answer Peter's first question below. You asked about finding shared segments without phasing which means you don't know which one out of the two segments you are sharing on. So that means three people can share with each other without all sharing on the exact same segment even though the segments line up in exactly the same places. Each person can match one person on the segment coming from his/her mother and also match the other person on the segment coming his/her father. I call it a three-way but not exact match. Example, John, Kate and Mary all match each other at approximately 7 cM on a chromosome browser, in the same spot and the segments are IBD. John and Kate match the German ancestor here. John and Mary match the Irish ancestor here. Kate and Mary match the English ancestor here. There are three ancestral lines. Each person only matches two out of the three. Mary may not have any German ancestry and does not descend from that ancestral couple yet she still matches the other two people. It is like holding hands in a circle with two other people. There are three different matching possibilities and you don't match where the two other people are holding hands. But you all connect with each other. I think this is very rare but can certainly happen in communities where the ancestors were all living in nearby circles (pun intended) but not all related to everybody else through exactly the same ancestors. Kathy J. -- > On 11/15/13, Peter J. Roberts<peterebay@yahoo.com> > wrote: > > If three or more people share a 7cM segment > and they all match each other on > that segment then what is > the likelihood they don't all descend > from the same ancestor or > ancestral couple? > What would be other exceptions > besides double cousins? > Thanks and sincerely, > Peter > Peter J. Roberts

Jim, that's interesting, I can't follow  your argument.  Whereas my presentation was clear to me.  Maybe you're talking about closer cousins; my writing was about single segments.  And single segment matches only match ~1/2 of our fifth cousins.  My analysis: I got 1/2 my mom's variable DNA (hereafter referred to as DNA).  I got 1/2 my dad's DNA.  Ditto my sis. The 1/2 from my mom in common with my sis is 1/2 *1/2 =1/4 my genome.  Ditto my dad. So I have 1/4 + 1/4 = 1/2 my DNA in common with my sis.  (All on average of course,  I actually have 48% in common with my sis.) Similarly the odds that we each got the same 1/2 from each parent is 1/2 *1/2 = 1/4 of the genome where we are 100% IDENTICAL.  So of the 1/2 the genome where I match my sis 1/4 / 1/2  = 1/2 of it tells me nothing new at all because we are identical there.  And 1/2 of the other spots we match, we'll match, therefore, corroborating our match, 1/2 we won't match, therefore,  saying nothing.  And 1/4 of our genomes, we won't match the same people because we don't even match each other. Summary: 1/4th genome, we are totally different, therefore, having totally different matches.   1/4th totally the same, therefore, no new info by our matching. 1/4 genome HIBD because we both match the same people, thus corroborating our match (the chances that our parent was HIBS and we each are also HIBS is too small to consider) 1/4 each of our genomes--either HIBD or HIBS because we match someone and our sibling doesn't because we each got the same DNA from our "wrong" parent who doesn't match the person we each match. Different expression of bottom line: At only 1/4th of our genome will we hit the jackpot because we are only HIBD with our sib and also HIBD with our matches.  AKA Corroborated. At 1/4th of our genome we'll totally understand why we don't match the same people because we don't match each other. At 1/4th of our genome, we'll match because we are 100% identical there rather than HIBD.  Therefore, no new info.  AKA Not Corroborated (not anti-Corroborated either, i.e either HIBD or HIBS with our putative match) and we have all the same matches. At the other 1/4, we might be corroborated and we might not but we have a lot of different matches from our sib. I don't know what's so difficult. Message: 3 Date: Sat, 16 Nov 2013 11:31:31 -0500 From: Jim Bartlett  Subject: Re: [AUTOSOMAL-DNA] > 1. Re: Finding shared atDNA ancestry without phasing (Jim Bartlett) To: "autosomal-dna@rootsweb.com"  Message-ID: <699613D6-572D-415E-BE77-42FD47033576@verizon.net> Content-Type: text/plain; charset=us-ascii I'm not sure how my partly true response relates to the rest of your email.  I will admit, however, that what I often say is partly true. I prefer to say it's mostly true;>j. DNA is random, so many outcomes are possible.  For your discussion about siblings - let's say the atDNA from a parent can be aggregated into 4 groups: a, b, c and d; so a+b+c+d=1. You got say a+b=.5; your sibling got b+c=.5; you both got b, neither got d. Do the algebra and a=c and b=d. So the amounts in each category can be determined. Note a and b are rarely equal, although over a large enough sample the average would be.  This aggregate analysis is not the same as a Triangulation analysis of segments among cousins. When you have segments among you and two Matches, on the same area of a chromosome, you have Triangulation (Y=A=B=Y), the segments are IBD - the chances that any of these segments are IBS are slim (almost 0). It's hard to make a concrete statement and waffle it just a little to allow for that random possibility.  Triangulation analysis is not usually done among siblings and very close relatives - for two reasons: 1. The relationships and Common Ancestors are known 2. A large "segment" between siblings and close relatives will involve multiple ancestors. You get a whole chromosome - a very large "segment" - from a parent, even a grandparent. Clearly there are multiple ancestors in that segment when it's broken up by matches with distant cousins. To a smaller extent this is true with large shared "segments" with aunts/uncles or close cousins. In fact this is the case with many matching segments - they come from multiple ancestors - but the smaller segments are below thresholds and we'll never see any cousin Matches on those more distant ancestors. The point is to be prepared to break large segments into smaller ones.  Jim - Sent from my iPhone - FaceTime! On Nov 16, 2013, at 8:59 AM, jlerch1@lighttube.net wrote: > Jim's statement about the chances being slim that 3 have an IBS is only  > partly true. My sister has recently gotten her data, and as expected  > at 1/2 of my matches at the 5th cousin level do I match her. But of  > that 1/2, 1/2 of that is at places where our parents each gave each of  > us the exact same DNA, so at that 1/4th of our genomes we are like  > identical twins. So we only count as 1 person at that 1/4th of our  > genomes. (It actually is a slight bit of corroboration since my data  > was V2 and hers is V3 chip.) You need to use the Family Traits app on  > 23 to see if you are Full IBD there instead of HIBD.  > 

On Family Tree DNA one can 'triangulate' for 'in common matches' (not necessarily in common on the same chromosome). But at least with that helpful hint I can quickly look at those people to see if they match on the same chromosome. Does 23andme have a similar tool? Or do folks all download the data from www.dnagedcom.com so they can manually review which people match on a specific chromosome? I am aware Family Inheritance Advanced lets you compare 3 people to another but that doesn't tell you who ELSE in the list might share that same chromosome location. I'm much more familiar with FamilyTree DNA so perhaps I've missed the tool on 23andme. Denise

I'm not sure how my partly true response relates to the rest of your email. I will admit, however, that what I often say is partly true. I prefer to say it's mostly true;>j. DNA is random, so many outcomes are possible. For your discussion about siblings - let's say the atDNA from a parent can be aggregated into 4 groups: a, b, c and d; so a+b+c+d=1. You got say a+b=.5; your sibling got b+c=.5; you both got b, neither got d. Do the algebra and a=c and b=d. So the amounts in each category can be determined. Note a and b are rarely equal, although over a large enough sample the average would be. This aggregate analysis is not the same as a Triangulation analysis of segments among cousins. When you have segments among you and two Matches, on the same area of a chromosome, you have Triangulation (Y=A=B=Y), the segments are IBD - the chances that any of these segments are IBS are slim (almost 0). It's hard to make a concrete statement and waffle it just a little to allow for that random possibility. Triangulation analysis is not usually done among siblings and very close relatives - for two reasons: 1. The relationships and Common Ancestors are known 2. A large "segment" between siblings and close relatives will involve multiple ancestors. You get a whole chromosome - a very large "segment" - from a parent, even a grandparent. Clearly there are multiple ancestors in that segment when it's broken up by matches with distant cousins. To a smaller extent this is true with large shared "segments" with aunts/uncles or close cousins. In fact this is the case with many matching segments - they come from multiple ancestors - but the smaller segments are below thresholds and we'll never see any cousin Matches on those more distant ancestors. The point is to be prepared to break large segments into smaller ones. Jim - Sent from my iPhone - FaceTime! On Nov 16, 2013, at 8:59 AM, jlerch1@lighttube.net wrote: > Jim's statement about the chances being slim that 3 have an IBS is only > partly true. My sister has recently gotten her data, and as expected > at 1/2 of my matches at the 5th cousin level do I match her. But of > that 1/2, 1/2 of that is at places where our parents each gave each of > us the exact same DNA, so at that 1/4th of our genomes we are like > identical twins. So we only count as 1 person at that 1/4th of our > genomes. (It actually is a slight bit of corroboration since my data > was V2 and hers is V3 chip.) You need to use the Family Traits app on > 23 to see if you are Full IBD there instead of HIBD. >