A New Frazer Tests for YDNA

It is big news when a Frazer tests for YDNA. YDNA is what is passed down from father to son only. So it is the perfect test for a family surname study like we have for Frazer of Roscommon, Ireland. The other reason that it is big news is that there are not that many male Frazers around to do the testing. We previously had 2 tested for YDNA. Now with Rick’s new test, we have three.

Frazer YDNA and Genealogy

I (and others) have pieced together a Frazer genealogy. It is not perfect, perhaps. That is why DNA testing is needed to affirm the work that has been done. Autosomal DNA is good, but the effectiveness tapers off as we go back more generations. Also autosomal DNA is not specific to a surname, so it could apply to any of hundreds of ancestors the further we go back in time. YDNA on the the other hand, is specific to one male line – in this case Frazer. There are two types of YDNA. One is STRs which Rick tested for. These are good, but there are some problems in interpreting them as I will explain later in the Blog. SNPs are the most exact YDNA test. The other two Frazers, Paul and Jonathan have tested for SNPs through the BigY Test.

Rick descends from George Frazer from Martinsburg, NY. George Frazer moved his family to Canada and left many descendants. I have more on that family here. [Edit: Rick’s cousin Pat replies, “It was Richard Patterson Frazer that brought his family to Canada from Martinsburg.   George was born there but came to Canada with his family as a child.”] Many years ago, I convinced Bill, one of those descendants,  that he and I were related based on genealogy. Even though I wasn’t 100% sure myself, I argued my point with Bill and he finally agreed that we were related. Later, I convinced Bill to take an autosomal DNA test which did show that he was related to my family and many other Roscommon, Ireland Frazer Families. Since then, others have tested and confirmed the relationships. Now we have Rick who has tested his YDNA proving a more certain link on the male Frazer line. So this has been a several year journey starting with traditional genealogy, followed up with autosomal DNA and YDNA testing. Here is the tree of the three male Frazers who have tested for YDNA:

This genealogy is the best that we could do with what we have. There are some uncertainties about it. We have other’s research with names but no references. We have the Elphin Census of 1747 and this includes a widow Mary Frazer. She is presumed to be the widow of Archibald Frazer born around 1690. There are two other Frazers in that Census who represent the Archibald and James Lines. Further, I have supposed that James Frazer b. about 1804 is my ancestor. However, this requires the marriage of two first cousins. This was not that unusual apparently back in the day. Those belonging to the Church of Ireland wanted to marry other Church or Ireland people. In fact, there were laws in place requiring that. However, those from the Church of Ireland were in the minority.

According to the chart above:

  • All the testers have the top Archibald as a common ancestor. Rick and Paul also have James Frazer b. about 1804 as their common ancestor.
  • Rick is 9 generations from Archibald b. about 1690. Paul is 8 generations away and Jonathan is 7 generations away from the first Archibald.
  • Paul and Rick are third cousins once removed
  • Jonathan is a 6th cousin once removed from Paul and 6th cousin twice removed from Rick.
  • Assuming that Rick, Paul and Jonathan were born around 1950, there would be about 260 years from the first Archibald to our three YDNA testers.

Rick’s New YDNA STR Results

Based on the above tree, we would expect that Rick would be more closely related to Paul than Jonathan. That is the case. Ricks closest match is to Paul with a Genetic Difference (GD) of 2. His GD to Jonathan is 3. The GD is the difference in STR mutations between two people. Here are all of Rick’s matches by STRs. He has more matches than Paul or Jonathan:

Here is how Rick’s STR matches compare with Paul’s and Jonathan’s. The numbers are GDs. So a larger GD is a more distant YDNA match. A blank means more than 7:

  • I put the matches together that matched all three Frazers.
  • The light blue highlighted names have taken the BigY Test
  • The SNP Tree below based on BigY SNPs shows that Grant is a closer match to Frazer than Hayes. This seems clear from the STR matches above also
  • The chart points out all the Grant matches. Half of the 10 Frazer STR matches are with Grants.
  • Rick is the only Frazer matching Hayes. This backs up the SNP tree below.

These are the same names that I have been tracking by SNPs. So that is good that the SNPs and the STRs agree with each other. Here is the new SNP tree that was developed by the R1a L664 Administrator Martin:

  • Patton is the most distantly related to Frazer and doesn’t show on any Frazer’s STR match list.
  • Grant shows as the closest match, then Hayes.
  • Chisum doesn’t show on the SNP map as Chisum has not taken the BigY Test.
  • Stuart doesn’t show yet, but he is in the the process of taking the BigY Test, so he will eventually show on the SNP Tree.
  • If Rick had taken the BigY test, he would have formed a branch under the Archibald line on the lower right of the tree. Rick and Paul would likely share some of the SNPs now shown as private SNPs for Paul and both Paul and Rick would then have their own new private SNPs.
  • Rick must share all the SNPs in the yellow box above the ca 1600 A.D date. These are the five brand new SNPs YP6489 through YP6493 that are right now Frazer only SNPs.

A Simple New Frazer STR Tree

A STR Tree for the Frazer should follow the genealogical tree and the SNP tree. Here are the STRs and where Jonathan, Paul and Rick differ:

Normally, the mode would be taken as the oldest STR. However, in the case of DYS391 I believe that based on comparison with other STR tests for Grant, Chisum, Whittaker and Stewart, that the 10 value of DYS391 is the older value. Likewise for CDY which changes more quickly than other STRs, I took 35-38 to be the older value. Prior to Rick’s results, I had this tree for Jonathan and Paul:

In this simple tree, Paul had all the STR changes since the 1690 Frazer ancestor and Jonathan had none. However, there is some guesswork in drawing this out. One rule is to make the chart as simple as possible, which is what I did. This means having the least amount of changes. It is sometimes called the rule of parsimony. Ambiguities in STRs is why some people prefer the SNP trees as there is little or no ambiguity in SNPs. Now note the above chart. Paul and Rick share 391 = 11 and CDYa = 35-40. Here is the new Frazer STR Tree with Rick included:

This shows:

  • Paul has his own STR mutation of 576 = 19 which happened sometime after 1804.
  • Likewise, Rick has his own STR mutation of 444 = 13 which defines Rick’s relatively recent branch.
  • The GD of 2 noted above between Paul and Rick are for markers 576 and 444.
  • The GD of 3 between Paul and Jonathan are for markers 576, 391 and CDYa.
  • The GD of 3 between Rick and Jonathan are for markers 444, 391 and CDYa.
  • The STR tree gets more difficult to draw further back from the present time. This is because more variables come into play such as parallel mutations and back mutations.
  • It would be possible to draw a STR Tree for the Grant, Hayes, Stewart, Chisum, Whittaker Tree, but I will leave that for another time.

Summary

  • The Big Y test for Paul and Jonathan have defined shared SNPs for the Frazer line and Private SNPs for the Archibald and James Lines
  • Ricks STR test has defined shared STRs for the line of James Frazer who was born about 1804. Any future testers from this line should be able to confirm descent from this line by testing for 391 = 11 which is a slower moving STR. CDYa would be a less reliable STR as it is a STR which changes much more often than most.
  • STR mutations that are unique to Rick and Paul cannot be dated precisely as they could have happened anytime after James Frazer of 1804 was born until Rick and Paul were born. That would be roughly between 1830 and 1950.
  • At the 67 STR testing level, on average, one STR change takes place about every 140 years. Jonathan shows no STR changes the way the chart was drawn. This would be less than average for him. Paul and Jonathan should have had 2 changes each, but each had 3 changes. This means they got the extra two STR changes that Jonathan should have had since 1690. It all averages out.
  • All this DNA testing makes me feel more close to my Canadian Frazer cousins.

 

A New R1a-YP5515 Tree: Patton, Hayes, Grant, and Frazer

Thanks to a Patton, Hayes, Grant and 2 Frazers that have had a BigY test, and the R1a-L664 Administrator, we now have a new SNP Tree. I was happy to hear that Martin, the L664 Administrator, fresh from a 6 week holiday in Australia, had come up with a new tree for YP5515.

First, the YP432 Tree

Technically, at FTDNA where the testing took place, Patton, Hayes, Grant and Frazer are still listed as YP432. FTDNA is a bit behind. Here is the YP432 Tree that Martin has recently updated:

I had mentioned previously that YFull, a business that analyzes these results had branching under YP431, but not under YP5515. The above tree by Martin now shows 5 branches under YP5515 vs. 6 under YP431 on the right side above. Here is YFull’s version of the tree. In their defense, they are coming out with a new YTree this month.

Dating the New Tree – Scandinavians and Scotsmen

The other good news about the new tree Martin developed is the dates. Before the new tree came out, we had YFulls date of a common YP432 ancestor of 2800 years ago. That was not very helpful. Martin’s tree has a date for YP432 which is about 1500 B.C which is even older by 700 years. I will be watching for the new YFull YTree to see if they have a date adjustment for YP432 and a new date for YP5515. Martin goes all the way up to 900-1600 A.D. for the SNPs that the two Frazers share. That is a huge difference from YP432 to the Frazers of 3100 years. Martin also points out that it was likely between 1400 B.C and 400 B.C. when our Scandinavian ancestor or ancestors decided to move to Scotland.

That means that the Frazers and their Patton, Hayes and Grant relatives are actually Scandinavian if we go back far enough. I suppose after a few hundred years of being in Scotland, no one could tell that they weren’t originally Scots. They were certainly in the area before the adoption of surnames. And when those surnames were adopted, they were Scots surnames. On the other hand, the YP431 Branch of YP432 appears to have remained in Scandinavia for the most part.

The YP5515 Tree

The YP5515 Tree is very new even though it dates back to 400-1400 B.C. And at this point there are only 5 BigY tested people on the tree, so we are part of an exclusive club:

Here I have taken out the ID’s for privacy for Patton, Hayes, Grant and Frazer. I did add the two Frazer lines: James and Archibald. This shows that Patton branched off quite early. The tree also appears to show that as late as 800 A.D. Hayes, Grant and Frazer had a common ancestor. Keep in mind that this would have been before the adoption of surnames. Roughly between 800 and 900 A.D. Grant branched off. The dating of the common ancestor of James and Archibald Frazer at 1600 is interesting as we believe the genealogy shows a common ancestor at around 1690 or so. The biggest change was in the SNPs of the two Frazers. They went from FTDNA’s official YP432 to YP432>YP5515>YP6479>YP6488>YP6489.

The above Tree should be taken as quite authoritative. However, it is possible that it could change by the addition of branches as others test for the new SNPs. It appears with all the Patton SNPs that there is quite a bit of room for additional branching. In my previous Blog, I had mentioned a Stewart and Chisolm. These matched by STRs, but had not taken the BigY or SNP tests. It would be possible for Stewart and/ór Chisolm to test and more precisely place themselves on the tree above (based on the SNPs they test positively or negatively for).

Another thing to note on Martin’s Tree is that there are now SNP names rather than just position numbers. For those SNPs that are shared most of them now have YP numbers. The private SNPs that are unshared, remain as position numbers.

Where My YP5515 Tree Went Wrong

When Martin came up with his tree, he noted that mine was mostly right. Here was my mistake:

I had a question mark by Patton as he had a bunch of missing SNPs that others had. Martin rightly had Patton descending from YP5515 on one side and Hayes, Grant and Frazer on the other being all YP6479. It is a subtle distinction, but an important one. Basically it means that Hayes, Grant and Frazer do not descend from the Patton Line, but they both have a common ancestor.

The other difference I notice is that I had the James Line with 4 private SNPs and Martin had them with only 3. Martin does not show the James Line as having the Private SNP #19995371. I checked our James Line tester against Grant and see that they both share that SNP, so that cannot be a private SNP for our James Line tester.

Frazers One Step Further

There is one other male Frazer on the Archibald Line who is in the process of testing YDNA for STRs. If I were to get a test method for each of the Archibald Line Private STRs and this tester were to test for each of these, we would be able to have one more branch on the Archibald Line with a named SNP defining that branch. Those two people have the common ancestor of James Frazer b. about 1803.

What’s Next?

Next we wait for the new YFull YTree. This will give a second opinion on Martin’s Tree. However, YFull does not have all the tested people that Martin has. After that, we wait for the STR results of the other Archibald Line person.

First YFull Analysis of a Frazer BigY Test

I have written quite a bit on Frazer BigY testing. To see my Blogs on the subject, they should be categorized under Frazer YDNA. Paul and Jonathan have both tested for the BigY. BigY finds all the SNPs of the YDNA of a male. These SNPs are both known and previously unknown SNP. These SNPs change roughly every 100 years. There are terminal SNPs, private SNPs and public SNPs, so it gets a bit confusing.

Frazer of North Roscommon, Ireland Terminal SNP

First, I forgot to mention what YFull is. YFull analyzes the Family Tree DNA (FTDNA) BigY test. For Haplogroups like R1a, which is what the Frazers are, they are good at analyzing the information and coming up with new SNPs. Sometimes, they are more up to date than FTDNA which seems ironic as it is FTDNA that did the test to begin with. For example, YFull has my cousin Paul as R-YP5515. FTDNA has Paul’s terminal SNP as R-YP432 which is older. This is how I have drawn it in previous Blogs:

The bottom line is that Frazer and Grant and perhaps Patton are all R-YP5515, but FTDNA doesn’t show it yet. YP5515 is in a box with many other SNPs, but YP5515 is the representative SNP for the group. That means that a bunch of SNPs are in that block, but we don’t know which one was formed first. At any rate, it is a point at which branching occurs. That is one good thing about SNPs. As they are unambiguous, they are good at forming SNP trees, which as basically ancient family (or clan) trees.

FTDNA’s Haplotree Vs YFull’s YTree

Here is how Frazer shows up on FTDNA’s Haplotree:

FTDNA only shows YP431 branching off, but not YP5515.

Here is YFull’s YTree:

In the tree above, Paul is indicated with a red “new”. The YTree tells me a lot more than the FTDNA Haplotree:

  • The R-YP5515 branch beneath R-YP432 is a lot smaller presently than the R-YP431 branch.
  • FTDNA is missing a lot of SNPs under the YP431 branch
  • The R-YP431 branch has a lot of people that appear to have Norwegian ancestors. From my previous Blogs, I have shown that the YP5515 branch appears to be from Scotland.
  • YFull is popular for it’s dating. YP431 shows a common ancestor at 1900 years before present. I assume that once Jonathan’s results are submitted to YFull, there will be new SNPs named and a new date for YP5515.

YFull’s Look At Private SNPs

Here we get into terminology. YFull calls these Novel SNPs, but when I choose them, they are listed as Private.  A simplified explanation is that a Novel SNP is an unnamed one and a Private one is one that is not shared with anyone else (at YFull). YFull gives novel SNPs a YFS number. Here is the list of Paul’s best quality Novel SNPs:

This shows that Paul has 42 Novel SNPs. What I don’t show is that there were 16 ambiguous SNPs. The other categories beyond ambiguous had no Novel SNPs. Those categories are Low Quality, One Reading and INDELs.

I can download the Novel SNPs to an Excel Sheet. Here are Paul’s SNPs that are not shared with Jonathan:

This shows that Paul has 4 Novel SNPs, what FTDNA calls Variants. In my tree above, I had that Paul had 5. When I checked Variant 28804880, I see that was a mistake, so I’ll take that one out.

This is a better result as now there is symmetry between the Archibald Line and the James Line. This shows exactly 4 new novel variants on the Archibald Line and the James Line since their common ancestor. 150 years is an average period for new SNPs, so that could mean 600 years to the common ancestor if these are all actually new SNPs. This could mean that the Frazer had more variants than average or that this R1a branch of YP5515 has more than average. The actual time to the common ancestor is about 260 years if we have done the genealogy correctly.

FTDNA VS YFull

YFull exists because FTDNA is not the best at analyzing its own BigY results. However, FTDNA has a few features that YFull does not have. With FTDNA, one can tell who the matches are. In YFull the matches are just by ID. With FTDNA, I can match by Variants which is useful. At Yfull I can only find matches by known SNPs.

Number of Novel variants

As above, YFull shows 42 Novel SNPs for Paul. This is close to what FTDNA shows. FTDNA shows that Paul shares 36 Novel Variants with Jonathan and that he has 4 unique Variants. That adds up to 40. That would give the impression that YFull is showing 2 more Novel Variants compared to FTDNA. But that is not the case. Actually, there is very little agreement at all.

This is the top of my spreadsheet that compares the Variants. Y5515 is in gold. YFull does not report any of the Variants in the YP5515 block or above. FTDNA does as it is out to date with its Variants. That is why there are so many blanks in the Paul YFull column. That is a difference of 21 SNPs to begin with. Going down the spreadsheet, the Variants that are listed as being the same in both FTDNA and YFull are shown in the Paul YFull column:

This shows that there are only 17 Variants out of 40-42 reported that are the same in both FTDNA and YFull. Below this are 25 Variants that YFull lists that FTDNA does not list. This is perhaps because FTDNA lists only the highest quality Variants. The 25 Variants that YFull lists includes Acceptable and Ambiguous Quality SNPs.

Next Steps

I have recently heard from Martin who is the L664 Administrator. He is back from a trip to Australia and working on getting these SNPs into a new tree. Hopefully Jonathan’s results will make their way to YFull also for analysis and help build out the YTree.