Saturday 25 December 2021

December 2021 Update on BigY Results

This blog post has a couple of purposes.  The main one is to show what we're learning about how branches of the Pike family fit together.  Another one is to showcase some examples of how results from the BigY test have been able to confirm the Pike ancestry of some people who don't currently carry the Pike surname.  Moreover, they had no reason to suspect that they even had Pike ancestry prior to doing DNA testing.  The results that I will highlight in this post involve Group 1 and Group 8 of the Pike DNA Project.

Before we get to that though, there is some administrative housekeeping that I need to mention.  When I created this blog I enabled an email subscription service so that people can be automatically notified when new blog posts are made.  To do that I used Google's FeedBurner service.  However, FeedBurner eliminated email subscriptions from its functionality in mid-2021.  So I have now linked the blog to a similar service provided by MailerLite and have transferred the list of subscribers.  My hope is that the transition is seamless.  However, as today's blog post will be my first live test with MailerLite, please be understanding if a hiccup occurs.  There's a slight chance that subscribers might receive notifications from both FeedBurner and MailerLite (that is, if FeedBurner is still operating beyond its stated end date).

In preparation for discussing some DNA results, let me initially say that the BigY-700 test is the most advanced Y-chromosome test offered by FamilyTreeDNA.  It analyses over 700 STR-based markers (including the 111 STR markers that we display on our project's webpage for Test Results) as well as approximately 22 million nucleotides at which SNP-based mutations might exist.  As a quick reminder about these acronyms, STRs are short tandem repeats and SNPs are single nucleotide polymorphisms.  For anybody who wants a more detailed introduction to the BigY test, Janine Cloud (Group Projects Manager for FamilyTreeDNA) gave a presentation at the RootsTech Conference that was held online in February 2021.  A recording of her talk is available at YouTube, at least for the time being.

This powerful test is enabling us to rebuild family trees, as we saw with some examples that were highlighted in our previous BigY update in December 2020.  And as we will see below, it can even do this in cases where there has been a change in surname that may not have been well documented.

Let's start with our project's Group 8 cluster.

Group 8

The origins of our project's Group 8 are currently a bit of a mystery to us.  Within the group we have multiple Pike lines that separately trace back to the 1700s before getting genealogically stuck in places such as Kentucky, South Carolina and Massachusetts.  Details of how these lines are connected are not known at this time.  Likewise, we do not yet know where in England (or elsewhere) earlier generations of the family lived.

We have not previously discussed BigY results within Group 8 in any detail, mostly because there weren't enough results to work with.  However, there are now five members of this cluster who have received BigY results.  These five results are for:

  • Edward (kit 200742) who descends from a Lewis Pike who was born in Massachusetts about 1747.
  • Randal (kit B645868) who descends from a Thomas Pike who was born in the 1760s and who was resident in Adair County, Kentucky in 1810.  Family lore says that Thomas came to the USA with a brother named Preston, but this has yet to be confirmed by historical records.
  • An anonymous Pike with kit number MK53973.  His kit is managed by his cousin Connie, who has provided genealogical details tracing back to a William Pike who was born about 1738 and who lived in Edgefield County, South Carolina.
  • Charles Ward (kit N38648) who descends from a Shadrack Ward who was born about 1765, likely in South Carolina.
  • JD Ward (kit 263255) who also descends from Shadrack, but via a different son than Charles.

Note that two of these five results are for men with the surname WARD and for whom there is no known PIKE ancestry.  These two Ward kits are managed by Charles' sister Melanie, who has done a lot of work in tracing the Ward family.  Traditional STR-based Y-DNA testing showed that her Wards carry a similar genetic profile as the Pikes of our Group 8.  Close matches like this with men having different surnames are not uncommon, and in many cases the connection seems to pre-date the advent of surnames.  But in this case there was the extra detail that the Y-DNA for Melanie's Wards did not match that of other Wards in the vicinity, prompting additional scrutiny.  As will be explained shortly, we now have compelling evidence that Melanie's ancestor Shadrack Ward, who was born over 250 years ago, actually descends from a Pike ancestor.

In terms of how the five BigY testers in Group 8 are related to one another, the following diagram illustrates what we knew in advance of their BigY test results.  The two Ward men (Charles and JD) are shown on the left.  Each square represents a man, and those in green have done the BigY test.

Note that aside from knowing that the two Ward men are fifth cousins once removed to one another, nothing else was known about the nature of any other relationships between these five men.

It is evident from their STR-based DNA results that there are connections.  But trying to determine the nature of connections from STR markers can be faulty.  And in this group of men we have a nice example.  Take a close look at the table of STR results for Group 8 and compare the 111 STR markers for Charles and JD.  You'll see that their results differ on 6 of these 111 markers.  These differences reflect mutations that arose in their lineages subsequent to when they split apart about five generations ago.  Because there are 6 differences, we say that their 111-marker "genetic distance" is 6.

Now take another look at the table of STR marker values and observe that JD and Julian (kit 143228) differ on only 4 markers.  So their 111-marker genetic distance is only 4.

Genetic distance is often used to estimate how near/far relationships are between men, with smaller genetic distances suggesting closer relationships.  However, these STR-based genetic distances can be misleading.  If we were to rely only on these genetic distances, we would falsely conclude that JD is more closely related to Julian than to his Ward kinsman Charles.  In this case it appears that several mutations arose along the line between Charles and his ancestor Shadrack, while Julian's Pike line appears to have had fewer mutations.  Moreover, one of the mutations in Julian's line looks to have occurred on the 94th marker, which also seems to have experienced a mutation within JD's line.  The simple calculation of genetic distance does not account for these multiple mutations on the same marker.

I'm not saying that we should completely ignore genetic distances.  Rather, we should not rely on them without understanding some of their potential pitfalls.

Let's now direct our attention away from the STR-based markers and towards the SNP-based markers that comprise the BigY test.  Mutations based on SNPs are much better measures of relatedness because unlike STRs, once a SNP mutation occurs it is essentially permanent for all subsequent generations.  This enables us to use these SNPs to determine how branches of a family tree fit onto that tree, even if we don't have accompanying genealogical records.  Let me highlight a point:  although STR-based matches indicate that two men are related in some way, SNP-based matches can provide insight into how they are related.

In the context of the five men from Group 8, the diagram shown earlier shows four separate branches of what must be a single larger family tree.  That much we know from their close STR-based results, and the genealogical work tracing Charles and JD back to their shared ancestor Shadrack.  But how these four branches fit together is the question we want to now consider.

For people who have taken the BigY test, FamilyTreeDNA provides a Block Tree that shows their place in the human Y-chromosome SNP tree.  Below is the portion of this tree that is shown for members of our Group 8.

To say a few words about how to interpret the information presented in this diagram, each block of SNPs (in which the SNP names are listed) acts like an umbrella in the sense that everybody shown underneath the block carries each of the SNPs listed for the block.  So JD, Charles, and Connie's cousin all share the SNPs named FT23630 and FT23646.  They also have each of the five SNPs in the higher block (with SNPs named BY142754 through to FT19630); Randal and Edward also carry these five SNPs.  However, Randal and Edward do not have either of the SNPs named FT23630 and FT23646.  Moving upward, next is a block with a single SNP (named BY79375) that has been inherited by all five of the men we're discussing as well as a non-Pike.  This non-Pike has not inherited any of the five SNPs in the block containing BY142754 through to FT19630.

The SNP listed as the first SNP within each block is prefixed with the haplogroup letter (in this case the haplogroup is I).  When multiple SNPs are listed together within a block, their order does not indicate the chronological order in which their mutations occurred, because that isn't yet known.  Only when somebody comes along with new test results showing they carry some SNPs of a block, but not others, can the relative order of SNPs be figured out.  And when that happens, FamilyTreeDNA revises the block tree to reflect the new discovery.

The block tree presented by FamilyTreeDNA allows us to rebuild the structure of the family tree for Group 8 as follows:

That is to say that the tree begins with the Pike founder for Group 8.  This tree then experienced a split, with Connie and the Wards belonging to one branch of this split.  It is along this line leading to Connie and the Wards that the two mutations FT23630 and FT23646 arose.  Edward and Randal (who were not reported to share any SNPs that the others lack) currently appear to descend from two other separate lines.  They do have SNPs of their own, but these SNPs are considered "private" and will only be added to the block tree when FamilyTreeDNA finds other people who also have them.

Note that because the Pike surname is common to all three of the main branches of this tree, the surname therefore pre-dates the branching, which is to say that all five men share a common Pike ancestor.  For the two Ward men, this is significant, as it proves that they descend from a Pike.  Moreover, since Charles and JD descend from two separate sons of Shadrack Ward, this proves that Shadrack himself descends from a Pike.

It is believed that Shadrack was born in South Carolina about 1765, not far away from where Connie's ancestor William Pike lived.  William is believed to have married about 1765, having come to South Carolina from England.  The obvious question to ask now is whether William was Shadrack's father, which would correspond to the dashed line in the above tree.  Prior to doing any DNA testing, this question was not one that anybody had any reason to wonder about.

To ponder this question, first note that JD, Charles and Connie's cousin are grouped together within the FamilyTreeDNA Block Tree.  Hypothetically, if a new SNP mutation had occurred with the birth of Shadrack, then it would have been passed on to each of JD and Charles (but not to Connie's cousin) and so FamilyTreeDNA would have used it to create a new SNP block for JD and Charles.  Since there is no such block, we can conclude that no new SNP mutations occurred with Shadrack's birth.  Likewise, there were no new SNP mutations along the portion of their ancestral line that goes from Shadrack back to the Pike ancestor they share with Connie.

BigY results that we've previously seen in the Pike DNA Project have suggested that a new SNP mutation tends to occur about every generation or two.  Mutations are random processes, so there's no specific formula for how frequently they occur.  Nevertheless, for a line of ancestors (in this case, from Shadrack back to the Pike ancestor shared with Connie) to not experience any SNP mutations, it is likely that the line is short in length.  This would be consistent with the theory that Shadrack may be a son of William.  It may also be the case that Shadrack was a nephew of William, as that too would result in a short ancestral line.

At present we cannot definitively answer the question of who Shadrack's father was.  But with additional BigY testing we might be able to find an answer.  We know that Shadrack and William both carried the two SNPs named FT23630 and FT23646.  If one or both of these SNPs can be shown to have arisen with William's birth, that would provide evidence to prove he was Shadrack's father.  So how do we determine if one or both of these two SNPs arose with William's birth?  That's simple:  just do a BigY test for a descendant of one of William's brothers.  Well, it's simple except for the challenge of determining who William's brothers were and then tracing their descendants.  It has been speculated that William had a brother Joseph who settled in Orangeburg County, South Carolina, as well as a brother Thomas who lived at Charleston but returned to England after the American Revolution.  Researching these other Pike lines is a matter of ongoing research.


Group 1

A year ago when Group 1's BigY results were last featured, there were then five test results to work with.  Four more members of the group have since done the BigY test, bringing the tally up to nine.  One of these nine is anonymous and we have no genealogical information for him.  For the remaining eight, the diagram below illustrates our initial knowledge of how they're related to one another based on historical records.

Four of them (Larry, Michael, Alan and Roger) are known to descend from John Pike from Whiteparish in Wiltshire who settled in Massachusetts in 1635.  Stephen descends from a different John Pike who was buried in 1681 in the parish of St Mary Aldermary in London and whose son Samuel settled in Pasquotank County, North Carolina.  Jonathan descends from a Simon Pike who married in 1751 at Stratfield Saye, Hampshire.  And Lori (who manages her father's DNA results) traces her Pike line back to a David Pike who married in 1828 at Ramsbury, Wiltshire and subsequently moved to Ohio.

Jay manages the DNA results for a cousin of his with the surname GUESS.  Jay and his cousin trace their Guess line back to a William Guess/Guest who was born in 1762 in Virginia.  William had several brothers, and descendants of theirs who have done Y-DNA STR testing have received results that match one another.  However, their results do not match those of Jay's cousin.  Meanwhile though, Jay's cousin does match with descendants of William who have done Y-DNA STR testing.  These results collectively indicate that there may be a genealogical error or a case of misattributed paternity concerning William.  In the following discussion we will see that he descends from a Pike.

The Block Tree provided by FamilyTreeDNA for our Group 1 is shown below.

Observe that all of the Pikes of Group 1 are located beneath the block containing the five SNPs named YP5465 through to YP6209.  Immediately below this block are two others, one for the pair of SNPs named YP5461 and YP5463, and another for the single SNP named FT427056.  These two blocks represent a major branching point in the family tree for our Group 1.  Some people may notice and wonder about the block tree's second column from the right, which I have noted to be empty.  Sometimes these unoccupied columns appear in the block tree, although I'm not sure why.  This one may just be a placeholder that is available in the event that somebody new comes along and is found to carry the five SNPs of the YP5465 block but none of the three of the SNPs that make up the two blocks beneath it.

The block tree presented by FamilyTreeDNA allows us to rebuild the structure of the family tree for Group 1 as follows:

The two SNPs named YP5462 and FT178399 are carried by the descendants of two different sons of John Pike from Whiteparish, and so we can conclude that John himself carried these two SNPs.  Jay's cousin does not carry either of these two mutations, which signals that he does not descend from John.

Meanwhile, on the right side of this tree we see that it has two distinctive branches, one of which has Stephen and the anonymous person.  At present it isn't clear whether this anonymous person descends from Stephen's ancestor John from London.  However, the block tree provided by FamilyTreeDNA shows that Stephen and the anonymous person share a block of four SNPs and thereafter they have an average of two additional private SNPs.  With only a few private SNPs, and with Stephen's ancestor John having lived about 10 generations ago, my suspicion is that the anonymous person does descend from John, as suggested by the dashed line in the tree shown above.

The rightmost branch in the tree contains Jonathan and Lori.  They share no SNPs that are more recent than FT427056 (which is also shared by Stephen and his anonymous kinsman).  In this case my suspicion is that Lori's ancestor David does not descend from Jonathan's ancestor Simon.  My reasoning here is that Simon lived about three generations before David, and if Simon was David's ancestor then there would be a strong chance that a SNP would have mutated along the line from FT427056 to Simon and it would then have been passed along to David and Lori's father.  Such a mutation would also be inherited by Jonathan and would therefore have been reported by FamilyTreeDNA as a shared SNP.  But no such SNP is reported.

Something worth pointing out is that the BigY results for Group 1 confirm that Jay's ancestor William Guest has Pike ancestry.  The SNP results have provided the evidence that embed this Guest line within the Pike family tree.  Were it not for these genealogical DNA test results, we would not have had any reason to wonder if William wasn't a biological child of his presumed parents (named William Guest senior and Susannah Adelaide Howard).  Currently we do not know whether William was adopted, or perhaps he was raised by his maternal grandparents (such as might be the case if his mother was unwed).  These questions will require further research.  One thing that stands out from the BigY testing is that whomever William's father was, he wasn't a descendant of John Pike from Whiteparish who settled in Massachusetts in 1635.  Nor was William's father a descendant of Samuel Pike who settled in Pasquotank County, North Carolina.  This suggests that there is a yet-to-be-identified Pike line that belongs to Group 1 and which was in Virginia by the 1760s.

In addition to revealing the existence of this unidentified Pike line, Jay's cousin's results have also helped to refine the structure of the Pike family tree.  If you look back to our project's BigY update from a year ago, you'll see the block tree at the time had same the block of five SNPs (i.e., YP5465 through to YP6209) that it still has now.  However, a year ago, there was then a lower block with four SNPs (i.e., YP5461, FT178399, YP5462, YP5463).  This block has now been broken apart.  Because Jay's cousin has YP5461 and YP5463, we now know that these two SNPs occurred before YP5462 and FT178399.  Said another way, the results obtained from Jay's cousin have allowed us to see an instance of branching in the family tree that wasn't visible a year ago.  The new results from Lori, Stephen and the anonymous project member have also allowed us to see some branching elsewhere in the family tree.

Let me now draw your attention to the table of STR results for Group 1 in which the standard set of 111 STR marker values is shown.  These results from Group 1 are currently subdivided within the table as follows:

  • Known descendants of John Pike who settled in Massachusetts in 1635, all of whom we note share a value of 25 on the second STR marker
  • People with a value of 25 on the second STR marker, but who aren't known to descend from this settler John
  • People with a value of 24 on the second STR marker

This second marker and its value of 24 versus 25 has appeared for a long time now to represent a deep split in the family tree.  Specifically, it has looked as though a mutation arose in which a child was born with a value of 24, whereas his father had a value of 25.  And then this child passed on the value of 24 to all of his male-line descendants, while the rest of the family tree carried on with the original value of 25.  Something similar to this may well have happened, but the BigY results that we now have are showing the actual situation to be more complicated.

Across the bottom of the tree diagram shown above I have inserted the value of the second STR marker for each of the nine people that have done the BigY test.  Notice that Lori's father has a value of 25, but the other people in the right half of the diagram have a value of 24.  The tree structure is validated by the SNP results reported by the BigY test and so we find that there must have been more than one instance of a mutation on the second STR marker.  That is, our previous notion of a single mutation on the second STR marker was incorrect.  This is one of the caveats of STR mutations, namely that STR mutations happen often enough that one mutation can later be reversed by a subsequent mutation on the same marker.  Such a reversal is what we now see to have happened.  Specifically, it appears that early in the history of the right side of the family tree the second STR marker mutated from a value of 25 to 24, and then within the last several generations of Lori's line it underwent another mutation, this time from a value of 24 to 25.  There are some other more nuanced possibilities that could account for what we see, which is to say that this updated theory may yet be found to have a flaw in it.  We will therefore adjust as necessary if/when presented with new information that doesn't fit the current understanding of this marker's history.

Group 1 is the largest family cluster within the Pike DNA Project, representing about a fifth of our project members.  The many members of this group who trace their lineages back to the English counties of Wiltshire, Hampshire and nearby indicate that the family has been in this area for several centuries.  As was the case with the inclusion of new results from Jay, Lori, Jonathan, etc., it will be interesting to discover more about this family's branches and origins as more members upgrade and obtain results from the BigY test.  It will also be good to identify more family lines and determine their place within the family tree.

Group 2

To say a few brief words about Group 2, there have been some new BigY results this year but I'm going to defer an update for a future blog post.  In part this is because there is currently a BigY test in progress for a member of Group 2, and I'm hopeful that another test or two will be ordered during the current year-end sale being held by FamilyTreeDNA.

I'm also hoping to recruit some more project members who I suspect belong to Group 2.  One particular line of interest is a Pike line that I've been able to trace from the late 1600s in Poole in Dorset, to the 1700s in Portsmouth in Hampshire, to the 1800s in London, the 1900s in Gravesend in Kent, and most recently in the county of Essex.  

Disclaimer and Acknowledgements

As with any genealogical research, the reasoning and conclusions that are outlined above are based on what information is available at the time.  New discoveries will be assessed and incorporated as they arise, and may on occasion require adjustment to previous conclusions.  That's one of the best features of genealogical DNA testing, for each new test result provides new information that can reinforce past conclusions, or sometimes it can focus scrutiny onto something that isn't quite right.

Our project and the discoveries that arise from it are only possible because of the collective efforts of many people, especially those who have contributed DNA samples and genealogical information.

Thursday 24 December 2020

December 2020 Update on BigY Results

In this blog post I want to provide a status report concerning results from BigY-700 test results within the Pike DNA Project.  The previous BigY update for our project was done two years ago, in December 2018.

To begin with a quick overview, the BigY-700 test is the most advanced Y-chromosome test offered by FamilyTreeDNA.  It analyses over 700 STR-based markers (including the 111 STR markers that we display on our project's webpage for Test Results) as well as approximately 22 million nucleotides at which SNP-based mutations might exist.  DNA profiles based on STR-based markers serve our project well by distinguishing genetic clusters of Pikes from each other, thereby allowing us to identify who belongs to each of our "Group 1", "Group 2", etc., clusters.  But we would like to do more.  In particular, we would like to rebuild our family trees, putting branches back into their proper place, along with determining how long ago various branches of a common tree separated from one another.  Thankfully, the SNP-based results from BigY tests are now enabling us to do just that.

In order to demonstrate this, I am going to focus this blog post on the BigY results for our project's Group 1 and Group 2 clusters.


Group 1

Within our project's Group 1, we now have a total of five BigY results.  Four of them are for descendants of John Pike who was married at Whiteparish, Wiltshire in 1612, and in 1635 settled in Massachusetts.  Of these four, three are for descendants of John's son John, while the fourth descends from John's son Robert.  The fifth member of Group 1 with BigY results traces his ancestry back to a Simon Pike who was married at Stratfield Saye, Hampshire in 1751.

The four results from known descendants of settler John Pike, combined with the previously established knowledge of their positions within the family tree, will help us to get a sense of how often SNP mutations can happen.  That is, we can estimate how many mutations are likely to arise per generation.  

Here is a mini family tree that shows the relationships between the five members of Group 1 with BigY results.  For this tree we are not yet taking into account their DNA results.

Note that at this point we do not know whether Jonathan's ancestor Simon is a descendant of John Pike.  It seems unlikely, given that Simon was in England when he married in 1751, but potentially one of John's grandsons or great-grandsons might have returned to England and had a family there.  We'll obtain a definitive answer once we consider the DNA results though.

Back in 2018 when I produced the previous BigY status update I had to spend a lot of time illustrating SNP-based trees that arose from people's individual results.  I am very happy to say that FamilyTreeDNA now does this on our behalf by maintaining (and regularly updating) a Y-DNA Block Tree.  Below is a portion of the Block Tree that encompasses our Group 1, which FamilyTreeDNA makes available to those who have done the BigY test.  I have editted this a little bit, for instance by adding the names of the people involved.



To discuss this Block Tree and how to interpret it, let's first hone in on Larry, Michael and Alan.  They are all located beneath a blue block that contains a SNP named FT77897.  This SNP is their most recent common SNP, which all three of them share.  But after that their lines split apart.  Also, this FT77897 SNP is not above Roger or Jonathan, so it is a SNP mutation that arose somewhere on the line of descent towards Larry, Michael and Alan after their lineage separated from those of Roger and Jonathan.  By looking back to the mini family tree shown above we can see that this FT77897 SNP must therefore have arisen in one of the five Pike males represented by the squares that are immediately below settler John and leading towards Larry, Michael and Alan.

Above Larry and Michael is a blue block containing the two SNPs named Y88233 and FT78318.  Larry and Michael share them both, but neither was detected in Alan's Y-DNA.  Hence they indicate another branch point in the tree, which we can see corresponds to the mini family tree that we have previously determined from genealogical records.  There is also a more recent split in the tree, where Larry's line separates from Michael's.  This split isn't explicitly represented in the Block Tree just yet.  My understanding is that the branches of the Block Tree are based on SNPs that are shared, and as Larry hasn't yet been found to share any SNPs that are distinct from Michael's (and vice-versa) then for the time being they (i.e., Larry and Michael) are shown together in the Block Tree.  But if/when another person on Larry's branch (or on Michael's) does the BigY test and is found to have a SNP that Larry has but Michael doesn't, then that will prompt an update to the Block Tree so that a new branch gets shown.  Note that the Block Tree does indicate that "private" variants (i.e., SNPs not yet found to be shared with others) are present, so such yet-to-be-seen-as-shared SNPs do exist.

Now look at the blue block of four SNPs that, like an umbrella, covers Larry, Michael, Alan and Roger.  The four SNPs in this block are named YP5461, FT178399, YP5462 and YP5463.  Since these SNPs are shared by each of Larry, Michael, Alan and Roger, they must have arisen among their common ancestors.  The earliest such ancestor is John Pike who moved from Wiltshire to Massachusetts in 1635.  We can therefore deduce that John carried all four of these SNPs.  But whether they arose first with him or first within one of his ancestors is not yet clear.

What is clear though is that Jonathan's BigY results show that he carries none of these four SNPs.  This proves that he does not descend from John.  Jonathan's line must trace back to some ancestor that is also one of John's ancestors, which is to say that following picture is what the family tree actually looks like, beginning with a common Pike ancestor:


Something profound warrants being pointed out here.  We have now been able to link Jonathan's branch of the family tree into its proper relative place on the tree, despite the fact that we can only trace his line back to the 1700s.  It is precisely because of BigY test results that we are able to perform this reconstruction of the family tree and determine its structure.  As more and more people do the BigY test, we will in turn be able to add more branches to where they belong.

Something else that we would like to do is to estimate the age of the Group 1 Pike family, by determining when the first two branches of it split apart.  To do that is a bit tricky, and I hasten to point out that what we get is an estimate for which we cannot reliably quantify the margin of error.

Nevertheless, to try to estimate the time frame of the first branching point in the family tree, recall that John has been found to carry four SNPs that are not in Jonathan's line.  The question now is how long did it take for these four SNPs to arise in John's line.  Mutation rates for Y-DNA SNPs are not firm.  That is, mutations don't occur with any assured regularity.  An analogy that I like to use is a row of slot machines at a casino.  Asking how many generations need to pass for four SNP mutations to arise is similar to asking how many slot machines need to be played in order for four of them to yield a jackpot.  Because of the random nature of mutations (and also the random nature of slot machines), there isn't a firm answer.  In our discussion above, we observed that in the first five men leading from John towards Larry, Michael and Alan there is a single SNP mutation (named FT77897), so here it took five generations for one SNP to mutate.  But then in the next three generations continuing towards Larry and Michael two mutations occurred (namely Y88233 and FT78318).  Clearly the rate at which mutations occur is not constant.  But with a combined three SNPs over eight generations, we can roughly estimate that approximately one mutation will happen every two generations.  I repeat, with emphasis, that this is an estimate.  And in this case it is based on an extremely small data set!  But working with this estimate anyway, with John having four SNPs that Jonathan doesn't, this estimate suggests that their most recent common Pike ancestor would have lived about eight generations prior to John.  And noting that John was born about 1572, if we also estimate that there are 25 years between generations, then that would suggest that Jonathan's branch of the family tree split away from that of John's around the year 1372, give or take a margin of error that is likely several decades in size.  Even with such a wide margin of error, it is clearly apparent that our Group 1 Pike family is an old one, for which the Pike surname was adopted many centuries ago.

One more thing to point out concerning the age of the Group 1 Pike family is that there is a blue block of five SNPs (named YP5465 to YP6209) under which all current BigY results in Group 1 can be found.  The closest non-Pike BigY result is not under this block, which is to say that this non-Pike does not have any of these five SNPs.  He does, however, share each of the five SNPs of the next block up (i.e., the block at the top of the diagram, containing the SNPs named YP4102 to YP5464).  So it is somewhere along the sequence of five SNPs named YP5465 to YP6209 that the Pike surname became fixed for our Group 1 cluster.  With only five SNPs in this sequence, it therefore appears that the split between Jonathan's line and that leading to John and his descendants took place within the first ten generations of the Pike surname (where this value of 10 is estimated based on mutations happening about once in every two generations).

Group 2

Within our project's Group 2, we now have a total of ten BigY results, all from men whose Pike ancestors lived in/near the town of Carbonear in eastern Newfoundland.  The earliest reference that I have yet found that shows a Pike present in Carbonear is dated 1681, although exactly when Pikes took up residence in Carbonear is unknown.  Throughout much of the 17th and 18th centuries it was common for large numbers of people to annually commute from southern England and Ireland and back again to work in the fishery in Newfoundland during the summer months.  Before the early 19th century Newfoundland records of births, etc., are rare.  The result is that when substantive records come into being in the 1800s we find a plethora of Pike families in and around Carbonear, without clear indications of how they might be related to one another.  Trying to sort them out is a lifelong quest of mine, and is one of the reasons why I established the Pike DNA Project in 2004 when I first learned that Y-DNA testing was available through FamilyTreeDNA.

Below is a mini family tree showing how the ten men of Group 2 who have BigY test results fit into six different branches.



Georgia and her cousin Tom have the deepest lineage that has so far been traced within Group 2.  The details of how their line traces back to Thomas Pike who was married at Poole in Dorset in 1680 are something that I've been working on writing up, but it isn't ready yet (I'll let people know when it is).  The rest of us, however, each become genealogically stuck at/near Carbonear in the 1700s or 1800s.  A few us do connect with others though.  For instance, Fred is a known fifth cousin of mine, who I located and arranged for a BigY test to be done.

Rodney, George and Kevin descend from three different sons of a Timothy Pike of Carbonear who was buried there in 1838 at age 76.  As I will explain, establishing that this is so has taken effort, for in some cases the connection to Timothy was not readily known.

Rodney's recent Pike ancestors have lived at Channel-Port aux Basques on the southwest coast of Newfoundland since the early 1840s when a different Timothy Pike of Carbonear settled there.  Voters lists from the 1830s at Carbonear show that there was a Timothy Pike Senior as well as Timothy Pike Junior, suggesting that Timothy junior was a son of Timothy senior. 

George traces his Pike ancestry back to Timothy William Pike who was born at Carbonear in 1838 to parents Thomas Pike and Frances (née Thistle) who married at nearby Harbour Grace in 1824.  Thomas appears in several voters lists from Carbonear as "Thomas of Timothy" and there is a corresponding baptism in 1809 for a child Thomas with parents Timothy and Mary Pike.

Recent generations of Kevin's ancestry have lived at Pinware on the coast of Labrador.  The remote nature of this coastline means that genealogical records are particularly hard to find, because baptisms, etc., could be recorded in registers of any number of travelling clergy.  For instance, the 1872 baptism of Kevin's great grandfather Mark Pike was eventually found (written in French) in the Drouin Collection of records located in Québec.  Mark's father Solomon was born in 1835 at Carbonear, to parents Edward and Ann Pike.  Edward died at Pinware in 1879, having moved there with his wife and children sometime around the 1840s or 1850s.  Because there were several Pikes at Carbonear named Edward, trying to identify which one was this Edward has not been easy.  One key piece of evidence in this case is that there are some entries in voters lists from 1832 to 1844 for "Edward of Timothy" and then an absence from 1847 onwards for such an Edward.  This disappearance of "Edward of Timothy" coincides well with when Edward would have relocated to Pinware.

Below is an image from the 1844 voters list, showing "Thomas of Timothy" and "Edward of Timothy" recorded next to each other.

Source:  1844 Voters List of Conception Bay, GN 43/7 Box 7 at The Rooms (Provincial Archives)

This evidence tying Rodney, George and Kevin back to Timothy senior at Carbonear is less compelling than we would ideally prefer, as more records would normally be relied upon.  But the reality is that we can only work with what is available to us, and in this case there is no baptism record for either Timothy junior or Edward (as is not uncommon for people born at Carbonear prior to 1820).  Thankfully we can supplement the historical records with BigY test results.  A portion of the Block Tree that encompasses our project's Group 2 is shown below.



Rodney, George and Kevin appear under a blue block with three SNPs (named BY111567, FT95073 and Y140924) indicating that each of these three SNP mutations was found in their BigY test results, but not in the results of the other testers.  So these SNPs arose within their branch of the family tree, after it split away from the rest of Group 2, and before their three individual branches split away from each other.  Also note that on average they each have two "private" variants (i.e., SNP mutations not yet found to be shared with others).  This is consistent with the previous discussion in which it appears that Rodney, George and Kevin descend, respectively, from the sons Timothy, Thomas and Edward of Timothy senior of Carbonear.

I confess that I'm telling this story backwards, having discussed the voters lists before the DNA results.  Chronologically, these events happened the other way around.  That is, it was actually because the BigY results showed that Rodney, George and Kevin are closely related that I was prompted to review a history of part of the Pike family of Carbonear that Gilbert Pike (who descends from "Thomas of Timothy") wrote in 1999.  And it was Gilbert's mention of the voters lists, and especially how they include an "Edward of Timothy" who disappears from Carbonear in the 1840s, that helped me make the connection to Timothy for Kevin and his Pike line.  To point out a profound observation here:  the BigY test results successfully rebuilt the family tree and showed the nature of the relationship between Rodney, George and Kevin without relying on historical records.  It was subsequent to the BigY testing that the historical records were used to corroborate the tree structure and to name the ancestors involved.

Regarding Angus, myself (David) and Fred, the Block Tree produced by FamilyTreeDNA has separated my father Angus and I from Fred.  Specifically, it shows me and my father beneath a blue block of five SNPs, and then above that there is a blue block of two SNPs that we and Fred are all underneath.  Hence the three of us have the two SNPs named BY23384 and BY23385, but Fred lacks the five SNPs (named BY23399 to BY23866) that were found within both mine and my father's BigY results.  It is because my father and I were found to share these five SNPs that FamilyTreeDNA has split us away from Fred when forming the Block Tree.  Since we have already traced our lines back to this split, in this case we see that these five mutations that my father and I share (but which Fred lacks) must have arisen in the span of the five generations from the split down to my father.  So here we have an example in which there is an average of one mutation per generation.  If such a high mutation rate were to persist throughout the Group 2 family tree then that would be a great benefit to us in terms of trying to rebuild our family tree on the basis of BigY test results, since each birth would be marked by a new mutation (but alas, this is probably too much to hope for).

Concerning Tom, Philip, a male cousin of Laverne, and Bob, they are currently collected together without a block of SNPs other than the large one (consisting of the fourteen SNPs named BY24054 to PH1079) that covers all of the BigY testers within Group 2.  This simply indicates that none of these four people have recent SNP mutations that have been found to be shared with others.  It is only when shared SNPs are found that the Block Tree presents people in a separate branch.  So at this point it is looking as though Group 2's BigY results indicate a family tree that has six separate branches, as was depicted at the beginning of our discussion for Group 2:  one branch each for Tom, Philip, Laverne's cousin and Bob, a fifth one for me, my father Angus and Fred, and then a sixth one that includes Rodney, George and Kevin.  

On the basis of the BigY results, these six branches appear to have arisen at about the same time.  If every birth were to give rise to a new mutation (which relies on a high mutation rate) then this would in turn suggest that these six lines correspond to six brothers who are sons of our most recent common ancestor.  I personally doubt that that is our reality though.  Instead, I suspect that the mutation rate is closer to one mutation every two generations, in which case these six branches could have arisen in men who weren't all brothers but who could be a combination of brothers and cousins.  More work, and more BigY test results will be needed to better refine and clarify the reality of our situation.  That said, it is interesting to note that Thomas Pike who married at Poole in 1680 had a son John who is known to have had six sons.

Meanwhile, something else to observe from the BigY test results is the block of 14 SNPs under which the ten BigY testers of Group 2 are found.  Somewhere along this sequence of 14 SNPs is when it appears that our patriline took on the Pike surname.  This can be deduced by observing that next to this block is a lineage that leads down to a non-Pike.  As for how old the Group 2 Pike family is, if we suppose an average mutation rate of two generations per mutation, then this would suggest that we acquired the surname sometime during the span of about 28 generations prior to the six identified Pike branches that appear to branch apart in the mid 1700s.  That's a horribly wide timespan, ranging from about the year 1000 to 1700.  For now that's the best that we can do, because for now our closest non-Pike match really just isn't very close at all.  And also it's partly because we haven't yet found any Group 2 Pikes who aren't from Newfoundland.

Genealogical Disclaimer

As with any genealogical research, the reasoning and conclusions that are outlined above are based on what information has been found to date.  New discoveries will be assessed and incorporated as they arise, and may on occasion require adjustment to conclusions.  That's one of the best features of genealogical DNA testing, for each new test result provides new information that can reinforce past conclusions, or sometimes it can focus scrutiny onto something that isn't quite right.

For those who have not yet done the BigY test

Part of my goal in concentrating on Groups 1 and 2 in this blog post was so that I could showcase some examples of how the BigY test and the Block Tree that it produces have the power to rebuild a family tree and to reveal how genealogically disconnected branches fit onto it in relation to one another.  Also, it is through our collective effort that this becomes possible, which is to say that more BigY results will lead to more and better revelations.  So I encourage everybody who has currently only done limited Y-DNA testing to consider upgrading.  Our project also welcomes financial donations from anybody who might want to help sponsor BigY tests for others.

Thursday 17 September 2020

An overview of new results from the past year

In this blog post I want to review seven STR-based Y-DNA results for members that have joined the Pike DNA Project since the previous overview that was sent to our mailing list in June 2019.  You'll note that I have been tardy in providing updates lately.  This past year has been a challenging one for me, owing largely to an ongoing family crisis.  The COVID-19 pandemic hasn't helped either.

Not to dwell on those matters though, let me now describe our project's new members' results.  As I've done before, I will discuss these in conjunction with the genetic groups of our project rather than in chronological order.  The numerical marker values for each person's results are shown on our project's page for Test Results (links that jump to the results of individual groups are embedded in the text below).


Group 1

Within our project's Group 1, we have a single new result, for Colin (kit number B637431) whose Pike lineage traces back to a Thomas Pike who lived in the town of Malmesbury in Wiltshire, where he died in 1751.  The earliest known record pertaining to Thomas is dated 09 August 1712 when he and his wife Elizabeth baptised their son Henry at Malmesbury.  In June of 1713 Thomas was admitted as a commoner/freeman of the Old Corporation of Malmesbury.

Thomas appears to have been the first Pike to reside at Malmesbury.  Where he came from is currently not known, but given the presence of a Pike family five miles away at Tetbury it was reasonable to wonder if that might be Thomas' place of origin.

As it happens, our project already had a member whose Pikes came from Tetbury, namely Robert (kit 194314).  When Colin's Y-DNA results became available, they did not match those of Robert.  However, they did match those of our project's Group 1, which has members whose ancestral origins are widespread.  Within the British Isles they are as far flung as Cork, London, and Derbyshire, but there is an emerging pattern whereby lots of people in the group trace back to Hampshire and nearby counties. Hence it appears that this cluster has its earliest origins in that area. The two earliest instances that we know of for Group 1 place it at Whiteparish Hampshire as well as at Newbury Berkshire, both in the 16th century.

What this means for Colin and his ancestor Thomas Pike is that it now seems unlikely that Thomas came from Tetbury.  We can also rule out connections with the various other Pikes lines whose Y-DNA does not fall within Group 1, such as the families in Devon, Dorset and Somerset that have so far joined our project.  Having found a genetic connection with Pike families to the south and east of Wiltshire, that is where Colin can now direct more of his attention.


Group 2

We have three new results that fit into Group 2, for which many members trace their Pike ancestry to eastern Newfoundland, and thence to Poole in Dorset.

One of them is for Fred (kit 910724) whose Pike lineage traces back to a Thomas Pike of Carbonear, Newfoundland.  This particular Thomas Pike is also my most distant known Pike ancestor, as shown in my lineage.  Fred and I descend from two different sons of Thomas' son Thomas:  Fred descends from Moses, while I descend from Henry, such that Fred and I are fifth cousins to each other.  So it came as no surprise that Fred's DNA results are a strong match with mine and the other members of Group 2.  

Note that Fred and I are not a perfect match on our results though.  On markers #51 and #81 Fred has values that have not yet been observed within any other members of our project's Group 2.  This indicates that these two markers experienced mutations somewhere along the portion of Fred's lineage that came after the portion shared with me.  In this case, since we know when Fred's line and mine separated, we can begin to isolate these mutations.  That is to say, we can conclude that they each occurred either with Fred, or his father, or grandfather, ... or 3x great grandfather Moses, but not before that (because before that Fred's line overlaps with those of me and my father Angus, who share the same values on these markers as the remainder of the members of Group 2).

The second new result in Group 2 is for Kevin (kit IN87489) whose father Patrick (kit 504850) had previously tested his Y-DNA.  Their Pike lineage traces back to an Edward Pike who was born about 1810, got married in 1829 at Harbour Grace in Newfoundland, and in 1879 was buried at Pinware (sometimes written as Pied Noir in early records) in Labrador.  Something that is particularly interesting about Kevin's results is that they are not identical to his father Patrick's results.  Looking at the chart of results shown for Group 2, we see that they differ on marker #49 where Patrick has a value of 22 (as do most members of our project) whereas Kevin has a value of 23.  This is an example, not just of a mutation happening, but one for which we can precisely pinpoint when it happened, namely with Kevin.  

Curiously, this value of 23 on marker #49 has also been found in the results for me and my father (kits 23996 and N21510).  And since my 5th cousin Fred who is mentioned above has a value of 22, it must be that this marker also mutated along my father's line, somewhere between him and his 2x great grandfather Henry.  What I want to point out here is that marker #49 has mutated on two separate occasions, in two different parts of the greater family tree of our Group 2.

Parallel mutations such as this can cause confusion when trying to interpret DNA results.  For instance, because Kevin, my father Angus, and I all share the same value of 23 on marker #49, it looks as though the three of us should belong close to each other on the family tree.  But that's a conclusion that relies on the mutation having happened only once, which in this case it did not.  Having test results from as many Pikes as possible helps to sort out situations like this.  So too does having results from the advanced BigY test, which I'll say a few words about towards the end of this blog post.

The last of the three new results for Group 2 is for Alex (kit IN69253), whose Pike/Pyke lineage also involves Labrador, and Carbonear before that.  Specifically, Alex traces back to a Nathaniel Pike who was born about 1809, likely at Carbonear where he married in 1837 and had three children baptised in the 1840s.  Nathaniel is one of a few Pikes from Carbonear who moved to Red Bay in Labrador at about this time.  Incidentally, Red Bay was awarded UNESCO World Heritage Status in 2013 in recognition of the whaling station established there by the Basques in the 1500s.

As yet we do not know how Alex and his ancestor Nathaniel are related to the other Pikes of our Group 2.  One interesting observation about Alex's DNA is that he has a distinctive value of 29 on the 21st of the Y-DNA markers.  Up until now we had only seen a value of 30 for this marker in our Group 2.  With just the one instance of the value of 29, we cannot yet say whether it is a recent mutation or if it has persisted for several generations of Alex's line.

Group 6

Peter (kit IN32180) has limited knowledge of his paternal ancestry.  His father Lawrence was born in 1913 in New Zealand to Irene Scottow Phillips, but who Lawrence's father was is not known.  Peter tested 111 STR-based markers and his strongest matches are with the members of our project's Group 6, some of whom match on as many as 107 of the 111 markers.  Although our project includes some Pike lines in New Zealand, we have not previously found any Pikes from Group 6 there.  So the task now is to try to find some branch of the Group 6 family (many of whom trace back to a James Pike who lived in Massachusetts in the 17th century) that was in New Zealand around 1912.


Group 7

John (kit number 929521) traces his Pike lineage back to a John Pike who was born in North Carolina around 1795 and died in Alabama in 1852.  Two other descendants of this ancestral John Pike have previously joined our project, and their results are a close match to those of newcomer John.  As we've seen in other examples, the match is not perfect.  In this case we see that John has a distinctive value of 18 on marker #32, whereas all other results within Group 7 show a value of 16.  Coupled with the knowledge of how John's line fits into the family tree, forming a branch with only three generations, we can determine that a mutation has taken place on marker #32 within the three most recent generations of John's line.

Incidentally, Group 7 also contains a member with an ancestor Phillip Pike who settled in Maine in the very early 1700s and who is believed to have come from Manchester in Lancashire.  As yet we do not know how John's line in North Carolina and Alabama is connected to this other line.


Group 10 

James (kit number 285064) descends from a William Peak who lived from 1825 to 1882 at County Down, Ireland.  James' DNA results strongly match those within our project's Group 10, many of whom have the surname McPike or a variant thereof.  Of those members of Group 10 who have traced their ancestry to the British Isles, so far all have traced back to Ulster.


The BigY Test

In the past year our project has also had several members receive results from the advanced BigY test, which analyses not just the standard 111 STR-based markers but also several million SNP-based markers on the male Y chromosome.  STR-based markers can occasionally have parallel mutations (as we noted above in Group 2) as well as back-mutations (such as when a mutation has the effect of reversing a mutation that occurred on the same marker a few generations earlier in a lineage), and these can make it difficult to reconstruct family trees.  SNP-based mutations are far less prone to these issues, and so they are a powerful tool in determining how the branches of a family tree belong together on the tree.

I'm very excited about the BigY test, its utility for showing the structure of family trees, and also the potential that it has for estimating ages of trees and their branches.  I will be showcasing the BigY test in a subsequent blog post.


Saturday 15 February 2020

Welcome to this blog

The Pike DNA Project was founded in July 2004.  Within a few months an email mailing list was established at so that project members and friends could receive periodic news and updates about the project and its discoveries.  An archive of all messages ever distributed via the mailing list can be found by clicking here.

In March 2020 all mailing lists at Rootsweb will be discontinued.  In order to maintain a means of sharing news about developments concerning the Pike DNA Project, this blog has been created.  Please feel welcome to subscribe in order to receive email notifications when new blog posts are made.

While my intent with this blog is to primarily discuss aspects of genealogical DNA testing as they pertain the Pike surname and its variants, from time to time I may also make posts that do not dwell directly on Pike DNA.  For instance, I may on occasion discuss some non-DNA aspects of Pike genealogy and family history.  Likewise I may occasionally comment on matters of genetic genealogy that are beyond the scope of just the Pike surname.

And to do just that, allow me to say that in November 2019 I made my first proper visit to Ireland when I attended the 9th Genetic Genealogy Ireland conference.  At this conference I gave a presentation regarding mitochondrial haplogroup H5a5, which has been found to occur with high frequency here in my home province of Newfoundland & Labrador, yet it is exceedingly rare elsewhere.  My presentation can be viewed online thanks to Legacy Family Tree Webinars.  Among the many excellent presentations that were delivered at this conference is one by Iain McDonald on the topic of the BigY-700 test offered by FamilyTreeDNA.  The BigY test is a powerful tool for surname studies such as the Pike DNA Project.  As more and more people take this test, I anticipate that there will be a continuing discussion within this Pike DNA blog about the test results and their implications regarding Pike family history.  On that note, please stay tuned, and encourage Pike friends and relatives to participate and ensure that their branches of the family tree are duly represented in the project.